I PROPOXYPHENE, NORPROPOXYPHENE, AND

PROPOXYPHENE, NORPROPOXYPHENE, AND PROADIFEN (SKF-525A) ARE

MECHANISM-BASED INHIBITORS OF CYP3A4, CYP3A5, AND CYP3A IN

HUMAN LIVER MICROSOMES

Anna Ruth Riley

Submitted to the faculty of the University Graduate School in partial fulfillment of the requirements for the degree Master of Science in the Department of Pharmacology and Toxicology, Indiana University

December 2008

Accepted by the Faculty of Indiana University, in partial fulfillment of the requirements for the degree of Master of Science.

______

Sherry F. Queener, Ph.D., Chair

______

David R. Jones, Ph.D.

Master’s Thesis Committee ______

David A. Flockhart, M.D., Ph.D.

______

Lynn R. Willis, Ph.D.

DEDICATION

Special thanks to my family, and friends (especially MK), and MJK

who helped me during this journey

iii

ACKNOWLEDGEMENTS

To my committee:

Sherry F. Queener

David R. Jones

David A. Flockhart

Lynn R. Willis

Eli Lilly and Company, for sponsorship

ABSTRACT

Anna Ruth Riley

PROPOXYPHENE, NORPROPOXYPHENE, AND PROADIFEN (SKF-525A) ARE

MECHANISM-BASED INHIBITORS OF CYP3A4, CYP3A5, AND CYP3A IN

HUMAN LIVER MICROSOMES

The purpose of this study is to determine if propoxyphene and norpropoxyphene are mechanism-based (irreversible) inhibitors of CYP3A, and to determine if propoxyphene and norpropoxyphene are reversible inhibitors of CYP3A. Mechanism- based inhibition is a type of irreversible inhibition that results from an inhibitor or its metabolite binding to an enzyme during drug metabolism, which renders the enzyme nonfunctional.

Propoxyphene is an analgesic that is frequently prescribed in the United States and Europe. It is metabolized by CYP3A enzymes, and is an irreversible inhibitor of

CYP3A4. The major metabolite of propoxyphene is norpropoxyphene, which has not been extensively studied for enzyme inhibition. Proadifen (SKF-525a) is not a marketed drug, but it is a known CYP inhibitor that is structurally similar to propoxyphene and norpropoxyphene. Propoxyphene, norpropoxyphene, and proadifen were characterized in these studies with CYP3A4(+b5), CYP3A5(+b5) and pooled human liver microsomes.

Time-dependent and concentration-dependent loss of activity of CYP3A was measured by formation of testosterone product. Propoxyphene and norpropoxyphene exhibited the greatest inhibition with CYP3A in human liver microsomes, followed by CYP3A4(+b5), and CYP3A5(+b5). Both compounds formed metabolic-inhibitor complexes with

CYP3A4(+b5) and CYP3A5(+b5), but not with human liver microsomes. Proadifen was

a more potent inhibitor of CYP3A4(+b5) than of human liver microsomes and

CYP3A5(+b5). The KI values of propoxyphene and CYP3A4(+b5) and human liver

microsomes fall within the range of reported therapeutic blood levels of propoxyphene,

with reversible inhibition constants (Ki values) above therapeutic blood concentrations

for propoxyphene and norpropoxyphene. The KI values of norpropoxyphene and

CYP3A4(+b5) and human liver microsomes are higher than most reported blood levels,

except for blood levels after repeated dosing of propoxyphene at high concentrations. The predicted change in the area under the plasma concentration versus time curve of an

orally administered CYP3A substrate with propoxyphene (AUC'po/AUCpo) was calculated

for common CYP3A substrates. The AUC'po/AUCpo ratios are four to twenty-five times

higher with co-administration of propoxyphene based on in vitro kinetic parameters.

Propoxyphene and norpropoxyphene may cause adverse events when chronically

administered at high doses and/or when co-administered with other CYP3A substrates.

Sherry F. Queener, Ph.D., Chair

TABLE OF CONTENTS

LIST OF TABLES……………………………………………………………...... ix

LIST OF SCHEMATICS………………………………..……………..……...... x

LIST OF EQUATIONS………………………………………..……………...... xi

LIST OF FIGURES………………………………..……………..………...... xii

ABBREVIATIONS……………………………………………….….……...... xiv

CHAPTER 1-INTRODUCTION………………………….…….……….…………….….1

Purpose of the study……………………………….…….…….………...... …...2

Background information on the compounds used in this study..………..….…...... 3

Background information on enzyme inhibition……….....…...…...…...…...... 5

Background information on enzymes used in this study.....…...…...…...... 7

CHAPTER 2-MATERIALS AND METHODS…….…….…...…...…...………...... …..11

Overview of methods used in the study………...... ……...... ……...……..….....11

Chemicals………………………………………………...... …...... 11

Enzymes………………………………………………...... ……...... 11

Mechanism-Based (Irreversible) Inhibition Experiments……...... …...... …...... 12

Reversible Inhibition Experiments………………….…...……...... 13

Metabolic-Intermediate Complex Formation Experiments…………...... …...... 14

Data Modeling for Inhibition….…………………………...... ….....…...... 15

Experimental Predictions of In Vivo Drug Interactions with Propoxyphene and

Norpropoxyphene……………………………...... 16

CHAPTER 3-RESULTS…………………………………...... …...…...... …...... 17

Chromatography Data…….……….…….…………………...…...... 17

vii

Irreversible Enzyme Inhibition Data………………….…………...... 17

Metabolic-Intermediate Complex Formation Data………………….…...... 19

Reversible Enzyme Inhibition Data…………..…………….…………...…...... 20

A Comparison of Irreversible and Reversible Enzyme Inhibition Data...... 20

CHAPTER 4-DISCUSSION.……….………………..…………...…...…...... 22

CHAPTER 5-CONCLUSION……...... ……………….....…...... 34

APPENDIX-SUMMARY OF STATISTICAL METHODS...... 76

REFERENCES…………………………………………...... 77

CURRICULUM VITAE

viii

LIST OF TABLES

Table 1. Chemical Structures of Propoxyphene, Methadone, Codeine, Norpropoxyphene

and Dinorpropoxyphene

Table 2. Proadifen (SKF-525a) and SKF-8742 Chemical Structures

Table 3. List of CYP3A4 and CYP3A5 Common Polymorphisms

Table 4. Summary of Kinetic Parameters for Enzyme Inactivation (Irreversible

Inhibition) with Propoxyphene, Norpropoxyphene, and Proadifen

Table 5. Reversible Inhibition Ki Values

Table 6. Reported Therapeutic (Total) Blood Levels of Propoxyphene and

Norpropoxyphene

Table 7. Predictions of Propoxyphene and Norproxyphene Interactions with Other

CYP3A Substrates (AUC'po/AUCpo)

LIST OF SCHEMATICS

Schematic I. Irreversible Inhibition

Schematic II. Irreversible Inhibition: Chemistry of Proposed Metabolic-Intermediate

Complex Formation

Schematic III. Reversible Inhibition

LIST OF EQUATIONS

Equation 1. Irreversible Equation for Enzyme Activity (at time (t)), Enzyme Activity at

Time (0), and Kobserved (Kobs) at Time (t)

Equation 2a and 2b. Irreversible Inhibition Equation for kobserved, kinact, and KI and

Enzyme Activity

Equation 3. Competitive Inhibition Equation

Equation 4. Noncompetitive Inhibition Equation

Equation 5. Uncompetitive Inhibition Equation

Equation 6. Calculating AUC'po/AUCpo Using Kinetic Parameters

LIST OF FIGURES

Figure 1. HPLC Chromatogram of Extracted Sample After Incubation with Recombinant

CYP

Figure 2. HPLC Chromatogram of Extracted Sample After Incubation with Human Liver

Microsomes

Figure 3. Propoxyphene and CYP3A4(+b5)-Percent Activity v. Pre-incubation Time.

Figure 4. Propoxyphene and CYP3A5(+b5)-Percent Activity v. Pre-incubation Time

Figure 5. Propoxyphene and Human Liver Microsomes-Percent Activity v. Pre-

incubation Time

Figure 6. Norpropoxyphene and CYP3A4(+b5)-Percent Activity v. Pre-incubation Time

Figure 7. Norpropoxyphene and CYP3A5(+b5)-Percent Activity v. Pre-incubation Time

Figure 8. Norpropoxyphene and Human Liver Microsomes-Percent Activity v. Pre-

incubation Time

Figure 9. Proadifen and CYP3A4(+b5)-Percent Activity v. Pre-incubation Time

Figure 10. Proadifen and CYP3A5(+b5)-Percent Activity v. Pre-incubation Time

Figure 11. Proadifen and Human Liver Microsomes-Percent Activity v. Pre-incubation

Time

Figure 12. kobs v. Inhibitor Concentration for CYP3A4(+b5) and Propoxyphene

Figure 13. Kobs v. Inhibitor Concentration for CYP3A5(+b5) and Propoxyphene

Figure 14. Kobs v. Inhibitor Concentration for Human Liver Microsomes and

Propoxyphene

Figure 15. Kobs v. Inhibitor Concentration for CYP3A4(+b5) and Norpropoxyphene

Figure 16. Kobs v. Inhibitor Concentration for CYP3A5(+b5) and Norpropoxyphene

xii

Figure 17. Kobs v. Inhibitor Concentration for Human Liver Microsomes and

Norpropoxyphene

Figure 18. Kobs v. Inhibitor Concentration for CYP3A4(+b5) and Proadifen

Figure 19. Kobs v. Inhibitor Concentration for CYP3A5(+b5) and Proadifen

Figure 20. Kobs v. Inhibitor Concentration for Human Liver Microsomes and Proadifen

Figure 21. Propoxyphene and CYP3A4(+b5)-Percent Activity Relative to Control

Figure 22. Propoxyphene and CYP3A5(+b5)-Percent Activity Relative to Control

Figure 23. Propoxyphene and Human Liver Microsomes-Percent Activity Relative to

Control

Figure 24. Norpropoxyphene and CYP3A4(+b5)-Percent Activity Relative to Control

Figure 25. Norpropoxyphene and CYP3A5(+b5)-Percent Activity Relative to Control

Figure 26. Norpropoxyphene and Human Liver Microsomes-Percent Activity Relative to

Control

Figure 27. Proadifen and CYP3A4(+b5)-Percent Activity Relative to Control

Figure 28. Proadifen and CYP3A5(+b5)-Percent Activity Relative to Control

Figure 29. Proadifen and Human Liver Microsomes-Percent Activity Relative to Control

Figure 30. Metabolic-Intermediate Complex Formation by Propoxyphene with

CYP3A4(+b5)

xiii

ABBREVIATIONS

ACN Acetonitrile

AIC Akaike’s Information Criteria

AUC'po/AUCpo Ratio of the area under the concentration versus time curve in the

presence of an inhibitor to AUC in the absence of inhibitor, both

drugs orally administered

CV Coefficient of Variation

CYP Cytochrome P450

CYP3A Cytochrome P450 3A Family

CYP3A4(+b5) Cytochrome P450 3A4 with coexpressed cytochrome b5

CYP3A5(+b5) Cytochrome P450 3A5 with co-expressed cytochrome b5

fm Total hepatic elimination of substrate

FG Intestinal wall bioavailability of a drug

F’G Intestinal wall bioavailability of a drug in the presence of inhibitor

HPLC High Performance Liquid Chromatography

i.v. Intravenous administration

Iu Average steady-state concentration of inhibitor in blood (µM)

-1 Kdeg Rate of enzyme degradation (min )

Kinact Maximal rate of enzyme inactivation in irreversible inhibition

experiments (min-1)

KI Substrate concentration at half Kinact (µM)

Ki Reversible inhibition constant analogous to Km (µM)

Km Substrate concentration at half Vmax (µM)

xiv

-1 Kobserved, Kobs Observed rate of inactivation (min )

po Per os, oral administration of a drug (literally “by mouth”)

SBC Schwartz Bayesian Criterion

SD Standard deviation of the mean

SE Standard error of the mean

Uv/vis Ultraviolet/visible (light)

-1 Vmax Maximal velocity of product formation (min ) v. Versus

WRSS Weighted Residual Sum of Squares

CHAPTER 1-INTRODUCTION

Propoxyphene, a commonly used analgesic, may inhibit a major metabolizing enzyme, CYP3A. CYP3A metabolizes an estimated 60% of marketed drugs and endobiotics. CYP3A metabolizes many different types of drugs across multiple chemical classes and therapeutic classifications. Inhibition of CYP3A can reduce and prevent metabolism of CYP3A substrates, including co-administered drugs and chronically administered propoxyphene. Inhibition of CYP3A can lead to higher blood levels of these

CYP3A substrates, which may cause an increase in reported side effects or adverse events. Some reported overdoses of propoxyphene may be attributed to mechanism-based inhibition of CYP3A. The purpose of this study was to determine if propoxyphene inhibits CYP3A, and, if so, to characterize the inhibition and determine to what extent the inhibition may affect patients who take the drug

Propoxyphene has been prescribed for over fifty years, but has been associated with numerous overdoses. Propoxyphene, like other opioids, has high abuse potential and is associated with higher rates of self-poisoning than other classes of drugs (Ng and

Alvear, 1993). Accidental and suicide deaths have been attributed to the drug and its metabolite, either alone or in combination with other medications, including alcohol, benzodiazepines, acetaminophen, and other painkillers (McBay, 1976). Many cases of accidental overdose have been in combination with alcohol or other pain medication due to additive central nervous system depression (Gram, 1979) or liver toxicity due to acetaminophen overdoses in combined medication (Sheen et al, 2002). Propoxyphene has been associated with 2100 reported accidental deaths (38.6% of total propoxyphene deaths) in the US from 1981 to 1999, and 7109 total US deaths from 1999 to 2006. It was

found as one of the top ten drugs identified during autopsies and implicated in 5.6% of drug related deaths from 1981-1999 (DAWN, 2003). Propoxyphene products were withdrawn from the market in the United Kingdom in 2006 due to high numbers of fatalities at approximately 400 deaths per year (Lister, 2005). Norpropoxyphene, the primary metabolite of propoxyphene, has been associated with cardiac deaths in patients, and has a long half-life that allows for accumulation in the body (Inturrisi et al, 1982;

Holland and Steinberg, 1979). Proadifen is structurally similar to propoxyphene and norpropoxyphene, but is not a marketed drug.

Previous reports of adverse events with propoxyphene administration may be attributed to mechanism-based inhibition of CYP3A. Mechanism-based inhibition occurs when a substrate of an enzyme inhibits it irreversibly during the catalytic cycle, and renders the enzyme permanently inactive. Mechanism-based inhibition can result in a reduction in the total amount of active enzyme. As the active enzyme pool is reduced, fewer substrates may be metabolized until new enzyme is synthesized. As a result, substrates accumulate resulting in elevated blood level concentrations represented by area under the curve, or AUC. AUCs may be above the desired therapeutic window of effectiveness and safety, resulting in toxic concentrations and overdoses.

Purpose of the study:

The focus of this research was to determine if propoxyphene, norporpoxyphene, and proadifen irreversibly inhibit CYP3A4(+b5), CYP3A5(+b5), and CYP3A in human liver microsomes in vitro. In vitro reversible inhibition of propoxyphene, norpropoxyphene, and proadifen with CYP3A4(+b5), CYP3A5(+b5) and CYP3A in human liver microsomes was also assessed. The second part of this research was to

determine if irreversible inhibition is due to the formation of a metabolic-intermediate complex. Metabolic-intermediate complex formation occurs when an inhibitor irreversibly binds to the CYP enzyme during its catalytic cycle, forming a covalent complex which is visible spectrophotometrically.

The in vitro inhibition data were examined comparing the potency of propoxyphene, norpropoxyphene, and proadifen for reversible and irreversible inhibition.

The inhibitors were also compared for potency with the different isozymes (CYP3A4 and

CYP3A5). The inhibitors were tested for metabolic-intermediate complex formation.

Finally, the in vitro inhibition data were analyzed in comparison to in vivo blood levels of propoxyphene and norpropoxyphene to assess the clinical significance of findings.

Background information on the compounds used in this study:

Propoxyphene is an analgesic that is frequently prescribed in the United States and Europe. It was the seventeenth highest-selling generic drug in the United States in

2006, and ranked thirty-fourth for total retail dollars spent in 2006 at $260 million (Drug

Topics, 2007). Marketed under the brand names of Darvon, Distalgesic, Co-proxamol1, and Darvocet, propoxyphene is administered alone or in combination with non-steroidal anti-inflammatory drugs and/or caffeine to gain synergistic pain relief (Beaver, 1988;

PDR, 2000).

Propoxyphene contains two chiral carbon atoms for two pairs of diasteriomers (α- d,l and β-d,l) (Somogyi et al, 2004) (see Table 1, p37). The β-d,l racemate is pharmacologically inactive (Nickander et al, 1984), but the α-d enantiomer

(dextropropoxyphene) has analgesic properties (Gruber, 1956) and the α-l enantiomer has

1 Co-proxamol was withdrawn from the market in the United Kingdom in January of 2005 due to drug related suicides and deaths (Lister, 2005).

antitussive properties (Miller et al, 1963). The drug product is composed of the α-d enantiomer in a hydrochloride or napsylate salt (AHFS, 2007). Dextropropoxyphene relieves pain by targeting opioid receptors in the brain, and is a weak opiate agonist

(Miller, 1970). Dextropropoxyphene has one-half to two-thirds the potency of codeine

(Gruber, 1977). Its mechanism of action is similar to other narcotic analgesics, such as methadone and codeine, and it shares similar alkylamine chemical structures with these other drugs (Somogyi et al, 2004; McMahon, 1961; Feinburg et al, 1976) (Tables 1 and

2, p37-38).

Propoxyphene is metabolized by Cytochrome P450 3A (CYP3A) to norpropoxyphene (Somogyi et al, 2004). Norpropoxyphene is not a marketed drug, although it has analgesic properties and even greater local anesthetic effects than propoxyphene (Nickander et al, 1984). Norpropoxyphene is further metabolized to dinorpropoxyphene (Nash et al, 1975). Other minor metabolites of propoxyphene and norpropoxyphene have been identified, but are not commercially available (McMahon et al, 1973; Nash et al, 1975).

Proadifen has a similar chemical structure and chemical properties to propoxyphene and norpropoxyphene (Table 2, p38). These three compounds are tertiary alkylamines of similar molecular weight. All three compounds are composed of a hydrocarbon chain with an ester, two phenyl substituents, and a di-methylated or di- ethylated nitrogen (Anders and Mannering, 1966; Somogyi et al, 2004). Unlike propoxyphene, proadifen was originally developed by Smith Kline and French, but it is not a marketed drug. Proadifen is a potassium channel blocker and a nicotinic acetylcholine receptor blocker (Anders and Mannering, 1966; Buening and Franklin,

1974). It is metabolized to the de-ethylated product SKF 8742-A (Buening et al, 1974).

Proadifen has been used as a research tool in animals and in vitro as a general inhibitor of

drug metabolism (Cook et al, 1954; Anders and Mannering, 1966; Buening and Franklin,

1974; Jones et al, 2007). Proadifen is a known general cytochrome P450 inhibitor

(Schenkman et al, 1972; Bensoussan et al, 1995). Proadifen forms a metabolic-inhibitor

complex with CYP3A4 (Jones et al, 2007), although its inhibition of CYP3A4 and

CYP3A5 has not been extensively characterized.

Background information on enzyme inhibition:

Enzyme inhibition can be categorized as reversible and irreversible2 (Lin, 1998).

In reversible inhibition, the inhibitor and enzyme bind non-covalently. When the enzyme

and inhibitor disassociate, the enzyme is still functional. In irreversible inhibition, the

enzyme-inhibitor bond is usually covalent, the enzyme has been chemically changed, and

the enzyme is no longer functional after binding to the inhibitor (Lin 1998).

An overview of irreversible inhibition is summarized in Schematic I (p44). As shown, the inhibitor can act as a substrate that is metabolized by the enzyme to a product

(i.e. metabolite), but the enzyme can also be inactivated by the inhibitor (Silverman,

1995). Irreversible inhibition common with Cytochrome P450 enzymes (CYPs) is termed mechanism-based inhibition (MBI), and occurs when a metabolite is reactive and binds to the heme or protein of the CYP that caused its formation (Lin, 1998). The metabolite, or product, binds the CYP enzyme covalently, removing it permanently from the active enzyme pool so that it can no longer metabolize drugs. The enzyme can only be

2 Some scientists support three types of inhibition including quasi-irreversible inhibition as the third type. Quasi-irreversible inhibition is similar to irreversible inhibition except that the inhibitor can become unbound from the enzyme in an in vitro setting chemically (with addition of potassium ferrocyanide) or using radiation (Lin, 1998).

replenished by new synthesis3. Mechanism-based inhibitors may form a metabolic-

inhibitor complex with CYPs that can be detected experimentally at an absorption

spectrum of approximately 450 nm (Murray, 1997)4. Metabolic-inhibitor complex

formation with the nitrogen atom of an inhibitor and the iron atom of a CYP molecule is

proposed in Schematic II (p44). Often alkylamines form a covalent complex with CYPs,

particularly with CYP3A (Bensoussan et al, 1995). There are several known examples of

pharmaceuticals that form metabolic-inhibitor complexes with CYP3A enzymes

including some macrolides (oleandomycin, erythromycin) and some protease inhibitors

(amprenavir, lopinavir, nelfinavir, ritonavir, and saquinavir) (Polasek and Miners, 2005;

Ernest et al, 2004). These irreversible inhibitors can reduce the amount of free enzyme available to metabolize other drugs, potentially causing adverse reactions.

During reversible inhibition, the inhibitor binds to the enzyme, the enzyme and inhibitor then disassociate, and the enzyme returns to the active enzyme pool. The inhibitor or metabolite of the inhibitor does not form a permanent complex with the enzyme as occurs in mechanism-based inhibition. As demonstrated by Schematic III

(p45), an enzyme (E) can bind with an inhibitor (I) to form the reversible enzyme inhibitor complex (EI), or the enzyme (E) can bind the substrate (S) to form the enzyme- substrate complex (ES) which can yield a product (metabolite), or return to free enzyme

(E) and free substrate (S). The Ki is the dissociation constant for reversible inhibition. Ki is a ratio of the amount of free enzyme and inhibitor ([E][I]) to the amount of enzyme and inhibitor complex ([EI]) (Stryer, 1996; Zubay, 1998).

3 The half life for CYPs in the body is 1-6 days (Dossing, 1983). 4 Also called an Iron II Metabolite Complex (Naritomi et al, 2004).

Background information on enzymes used in this study:

CYPs are heme-based membrane proteins found in bacteria and animals.

Eukaryotic CYPs are 480 to 560 amino acids long, and can be found in the endoplasmic reticulum, mitochondria, or cytosol of the cell. Microsomes are self-sealing fragments of endoplasmic reticulum membranes (Stryer, 1996), and the CYPs found in the endoplasmic reticulum of cells are referred to as the microsomal type (Nelson et al,

1996). Microsomal CYPs were the focus of this study. CYPs are named for their absorbance peak at 450 nm, (Danielson, 2002). CYP enzymes can be found in the liver

(the major site of metabolism) and small intestine, kidney, skin, brain, lungs, gonads, adrenal glands, and other tissues (Goodman and Gillman, 1996). CYPs often metabolize highly lipophilic drugs into more hydrophilic compounds that can then be more readily eliminated from the body in the urine (Danielson, 2002; Goodman and Gillman, 1996).

CYPs require reduced nicotinamide adenine dinucleotide phosphate (NADPH)

CYP reductase, NADPH, and molecular oxygen to perform their oxidation/reduction reactions to metabolize drugs. NADPH reductase is a membrane protein located near the

CYP that contributes electrons to the oxidation reaction, NADP(H) serves as a cofactor by donating electrons to the reductase, and molecular oxygen can bind the ferric iron of the hemoprotein and eventually combine with the leaving group (e.g. N- or O-

Dealkylations reactions) or bind to the parent drug (N- or S-oxidations) (Goodman and

Gillman, 1996). Although they exist in vivo, CYPs and supporting enzymes, and cofactors (CYP reductase, and cytochrome b5, another electron donating group5), have

5 The proposed mechanism of reduced cytochrome b5 in CYP oxidation-reduction reactions is to transfer electrons to P450 after being reduced by NADPH-P450 reductase (Yamazaki et al, 1996). It can interact synergistically to boost catalytic efficiency of CYPs (Danielson, 2002).

been recombinantly expressed in bacculovirus cells for in vitro testing (Crespi and

Penman, 1997).

Thirty different cytochrome P450 enzymes have been identified including twelve families of CYPs in humans (Williams et al, 2004). CYPs with greater than 40% amino acid identity belong to the same family (Danielson, 2002). Of all the human CYPs, the

CYP3A family is one of the most important for drug metabolism. The CYP3A family in humans metabolizes endobiotics such as testosterone and progesterone (Niwa et al, 1998) and approximately sixty percent or more of marketed drugs (Wrighton et al, 1990), including propoxyphene (Chow et al, 2006).

CYP3A performs N-demethylation and hydroxylation reactions. Propoxyphene and norpropoxyphene are metabolized by CYP3A in the liver to norpropoxyphene and dinorpropoxyphene, respectively (Somogyi et al, McMahon et al, 1973; Nash et al,

1975). Compared to other CYP families, the CYP3A family metabolizes drugs with the largest molecular size (Nagata and Yamazol, 2002). CYP3A is the most abundant CYP in the liver and small intestine (Yamazaki et al, 1996); its expression level is thirty to sixty percent of the total CYP content in the human liver (Shimada et al, 1994) and comprises sixty to seventy percent of the CYP in the small intestine (Anttila, 1997). The enzymes

CYP3A4, CYP3A5, CYP3A7, and CYP3A43 make up the CYP3A family in humans

(Wrighton et al, 2000).

CYP3A4 and CYP3A5 share an eighty-three percent amino acid sequence identity (Aoyama et al, 1989). These two enzymes have similar substrate specificity (Lin et al, 2002), with CYP3A5 generally exhibiting a lower metabolic capability than

CYP3A4 (Williams et al, 2002). Most of the general population expresses CYP3A4

despite the existence of genetic polymorphisms. Keshava et al, (2004) found that polymorphisms in CYP3A4 do not contribute to differences in activity. Although Dai et al, (2001) found three genotypes associated with differential CYP3A4 expression, *17,

*18, and *1B. The CYP3A4*17 genotype corresponds with reduced CYP3A4 expression, and the CYP3A4*18 corresponds with increased CYP3A4 expression as compared with wildtype *1 (Dai et al, 2001), (Table 3, p39). Contradictory results have been reported for CYP3A4*1B. CYP3A4*1B expression is associated with increased CYP3A4 activity over wildtype (Kuehl, 2001). CYP3A4*1B is associated with reduced CYP3A4 expression (Wojnoski et al, 2002).CYP3A4 is generally considered the most abundant

CYP in the liver and the small intestine (Shimada et al, 1994). Unlike CYP3A4, CYP3A5 is expressed in the kidney (Eichelman and Burk, 2001), and is the predominant CYP in the lung (Attila, 1997).

CYP3A5 is polymorphic and includes individuals who do not produce functional

CYP3A5 enzyme. CYP3A5 is only detectable in twenty to thirty percent of human livers

(Eichelbaum and Burk, 2001). Ten to thirty percent of Caucasians, fifty-five to seventy percent of Black Africans and African Americans, and thirty-three percent of the

Japanese express CYP3A5 (Kamden et al, 2005). Several alleles have been identified for

CYP3A5; *1, *2, *3, *5, *6, and *7. The *1 allele is the wildtype allele. The presence of one *1 allele contributes to high expression of CYP3A5. CYP3A5*1 produces ten to thirty times the amount of CYP protein produced from CYP*3/*3 (Kreutz et al, 2005). In human liver microsomes, Huang et al, (2004) found that individuals carrying the

CYP3A5*1 allele, CYP3A5 constituted more than 50% of the total CYP3A expression.

Polymorphisms in CYP3A5 can be a causal factor in differential patient responses to

drugs and food products. Chow et al, (2006) found that poor CYP3A5 expressers

(CYP3A5*3/*3) had higher propoxyphene plasma concentrations and lower clearance rates of propoxyphene than expressers (CYP3A5*1/*3 or CYP3A5*1/*1). CYP3A7 and

CYP3A43 are enzymes that have been found in very low levels in some adult livers

(Gellner et. al., 2001), and CYP3A7 is predominantly a fetal enzyme (Thummel and

Wilkinson, 1998). For these reasons, only the CYP3A4 and CYP3A5 enzymes of the

CYP3A family are examined along with pooled human liver microsomes.

CHAPTER 2-MATERIALS AND METHODS

An overview of methods used in this study:

Propoxyphene, norpropoxyphene, and proadifen were incubated with recombinant

CYP3A4 and recombinant CYP3A5 recombinant or pooled human liver microsomes.

Formation of 6β hydroxy testosterone, the major metabolite of testosterone, was quantified with HPLC. Kinetic parameters were estimated from data fit with Windows

NonLin non-linear regression data models (version 5.0.1, Pharsight, Mountain View,

California). Kinetic parameters calculated from the experimental inhibition data were used to estimate in vivo changes in area under the plasma concentration time curve, AUC.

Metabolite complexes were detected by ultraviolet/visible spectrophotometry, based upon an absorption maximum of 450 nm.

Chemicals:

Propoxyphene hydrochloride was obtained from the United States Pharmacopeia

(Rockville, Maryland). Norpropoxyphene HCl, testosterone, 6β-hydroxytestosterone, desmethyl diazepam, temazepam, ammonium acetate, and NADPH were purchased from

Sigma-Aldrich (St. Louis, Missouri). HPLC grade acetonitrile (ACN) and methanol were purchased from J.T. Baker (Phillipsburg, New Jersey), and HPLC grade ethyl acetate was purchased from EMD Chemicals, Inc. (Gibbstown, New Jersey).

Enzymes:

Recombinant CYP3A4(+b5) and recombinant CYP3A5(+b5) were purchased from BD Gentest (Woburn, Massachusetts). Adult human liver microsomes were prepared from human liver tissues (in accordance with protocols approved by the

Institutional Review Board of IUPUI ). The homogenates of five livers were pooled at 20

mg/ml total protein yielding a CYP concentration of 0.3 nmol/mg protein. (Lowery et al,

1951; Gorski et al, 1994).

Mechanism-Based (Irreversible) Inhibition Experiments:

Enzyme inhibition was determined by the time-dependent and concentration- dependent loss of 6β-hydroxy-testosterone product formed from testosterone by

CYP3A4(+b5), CYP3A5(+b5), or human liver microsomes. The inhibitor was diluted in methanol and subsequently diluted in phosphate buffer (with 5 mM MgCl2, pH 7.4) for a final concentration of < 0.5 % methanol or evaporated to dryness prior to incubation with human liver microsomes (100 µg total protein), CYP3A4(+b5) (20 pmoles), or

CYP3A5(+b5) (20 pmoles). The reaction was started with 5 µl of 10 mM NADPH made with phosphate buffer for a final pre-incubation6 reaction volume of 50 µl at 37 °C in a mixing waterbath. Inhibitor concentrations and inhibitor incubation times were determined from preliminary experiments of single data points. Final pre-incubation experiments were conducted with samples in duplicate.

Immediately following the pre-incubation of inhibitor and enzyme, 950 µL of incubation supplement mixture (containing phosphate buffer as described above, 250

µM7 testosterone (substrate), and 10 mM NADPH) were added to the tubes in 37 °C water bath for an additional two minute incubation. The reaction was quenched with two mL ice-cold ACN. The internal standard, desmethyl diazepam or temazepam (600 ng), was added and the reaction tubes were mixed on a vortex mixer for 30 seconds and centrifuged at 2500 rpm for 5 minutes. The supernatant was removed from the protein

6 Pre-incubation refers to the incubation reaction of the inhibitor, enzyme, and NADPH, prior to the addition of substrate. 7 The concentration of testosterone was well above the Vmax for 6β hydroxyl testosterone formation with CYP3A4(+b5), CYP3A5(+b5), and HLM.

and added to glass screw-top tubes containing 5 ml of ethyl acetate. The tubes were shaken for 30 minutes and then centrifuged at 2500 rpm for 5 minutes. The organic layer was removed and transferred to 13x100 mm glass culture tubes and evaporated to dryness. That residue was reconstituted with 250 µl mobile phase (40% 30mM

Ammonium acetate, pH 6.4: 60% methanol (volume per volume)). Concentrations of 6β- hydroxy testosterone were determined using HPLC with 5 µm C-18(2) Luna

Phenomenex (Torrance, California) column with a 1 ml/min flow rate and uv detection at

254 nm.

Reversible Inhibition Experiments:

To test for reversible inhibition, inhibitor and testosterone were diluted in methanol and subsequently diluted in phosphate buffer (with 5 mM MgCl2, pH 7.4), for a final concentration of < 0.5 % methanol, or evaporated to dryness in tubes prior to incubation. Enzyme, substrate, inhibitor and phosphate buffer were combined for a total volume of 900 µL, and 100 µL of 10 mM NADPH solution were added to start the reaction. The tubes were incubated for two minutes in a 37 °C mixing water bath. The reaction was quenched with two mL ice-cold ACN and subsequent extraction and analysis steps were performed (as described above). The concentration of substrate and inhibitor used to calculate Ki values were estimated from preliminary experiments with single samples (see Data Modeling section below). Final experiments were conducted with samples in duplicate. A minimum of four concentrations of testosterone were tested, two concentrations above the Km and two concentrations below the Km for each enzyme.

Each inhibitor was tested with at least two concentrations above and two concentrations below the estimated Ki value (from preliminary data).

Metabolic-Inhibitor Complex Formation Experiments:

Initial metabolic-inhibitor complex experiments were conducted with a dual beam spectrophotometer (Uvicon 933, Research Instruments). Two 1 mL cuvettes were prepared with 100-200µL enzyme (100-200 pmoles CYP3A4(+b5) or CYP3A5(+b5) and

1 mg of total protein of human liver microsomes), 5 µL of 1 mg/mL8 inhibitor in methanol (or methanol for reference), and 0.1 M phosphate buffer with 5 mM magnesium chloride each at 37 °C. 100 µL of 10 mM NADPH were added to start the reaction (1 mL total reaction volume). Wavelengths 380-500 nm were scanned at time zero and then at two or five minute intervals, the samples were assessed for an absorbance maximum by subtracting the absorbance at 490 nm from the difference of the absorbance scan at each timepoint and a background absorbance scan. The metabolic-inhibitor complex forms as a characteristic peak at approximately 450 nm that increases with absorbance over time.

The method to detect metabolic-inhibitor complex was adapted for a microplate reader

(Synergy™ 2, BioTek) for follow-up experiments. In these experiments, 80 µL of enzyme and 260 µL of phosphate buffer (as described above) were combined in a polypropylene tube. Next, 170 µL of enzyme and buffer mixture were aliquoted into each well (experimental and control well). The plate was warmed at 37 °C for two to three minutes, and then 10 µL of inhibitor (in a 1 mg/mL solution of 10% methanol and 90% phosphate buffer) were added to the first well, and 10 µL of 10% methanol 90% phosphate buffer were added to the reference well. The reaction started with the addition of 20 µL 10 mM NADPH (Pershing and Franklin, 1982).

8 This corresponds to a concentration of approximately 15µM for each.

Data Modeling for Inhibition:

The percentage of enzyme activity remaining after incubation was determined by the amount of 6β-hydroxy testosterone formed relative to the amount formed at 0 time

(100%) at each inhibitor concentration. The natural logarithm of percent activity was plotted against pre-incubation time at different inhibitor concentrations. The lines of best fit were determined using Microsoft Excel best fit trendline. The slopes of these lines were used to determine kobserved values, or the observed rate of 6β-hydroxy testosterone product decline, at various inhibitor concentrations. Inhibitor pre-incubation time and percent activity (Et) at various concentrations of inhibitor were modeled with Windows

Nonlin Professional to estimate kinact, the rate constant for maximal rate of inactivation, and KI, the concentration of substrate at half maximal inactivation (Equation 1, p46).

From the Windows Nonlin Professional KI and kinact estimates, kobserved was calculated and plotted as a hyperbolic curve (see Equations 2a and 2b). The standard error of the mean (SE), coefficient of variation (CV), Akaike Information Criterion (AIC), and

Schwartz Bayesian Criterion (SBC) were used to evaluate each model for best fit.

Information on the statistical criteria can be found in the Appendix.

For reversible inhibition, 6β-hydroxy testosterone/min product formation data were modeled with Windows Nonlin Professional. The models estimated Vmax, Km, and

Ki values. The Vmax is the maximal velocity of product formation that can be achieved by increasing substrate concentration under the conditions of assay. Preliminary reversible and irreversible inhibition experiments were conducted using singlet data points across multiple concentrations of inhibitor. Experiments were conducted in duplicate and data were modeled in competitive, noncompetitive, and uncompetitive inhibition models (see

Equations 3-5, p46). Inhibition type was determined based on best fit criteria from model

(lowest SE9, CV, AIC, SBC). Final experiments were conducted in duplicate using a minimum of two concentrations below and two concentrations above the estimated Ki value.

Experimental Predictions of In Vivo Drug Interactions with Propoxyphene and

Norpropoxyphene:

The in vivo AUC'po/AUCpo of common CYP3A substrates was estimated using

Equation 6 (p46) (Wang et al, 2005). FG is the intestinal wall bioavailability (Wang,

2005), and was determined from published data. F’G is the intestinal wall bioavailability in the presence of the inhibitor. The kinact and KI values were determined from experiments with human liver microsomes. Iu is the average unbound steady state concentration of each inhibitor. The kdeg values, or rate of enzyme degradation, are

0.00128 and 0.00026 min-1 based on rat CYP3A and human CYP3A4 in CaCO-2 cells

(Correia, 1991; Malhotra et al, 2001). Both rates of enzyme degradation were used in the calculation resulting in a range estimates for the AUC'po/AUCpo. The equation assumes a maximal inhibition of CYP3A by propoxyphene or norpropoxyphene from the gut wall, and F’G is equal to 1 (consistent with Wang et al, 2005 and Ernest et al, 2004). The total

CYP3A hepatic elimination of the substrate without inhibitor (fm) was determined from published in vivo data.

9 See Appendix for more information on statistical criteria used to evaluate inhibition models.

CHAPTER 3-RESULTS

Chromatography Data:

For incubations with CYP3A4(+b5) or CYP3A5(+b5), a mobile phase of 40% 30 mM ammonium acetate (pH 6.3-6.4) was used with desmethyldiazepam or temazepam as internal standards. Non-extracted standards (6β-hydroxy testosterone, internal standard, testosterone), and inhibitors were analyzed by HPLC prior to use in an incubation to verify peak separation and recovery (Figure 1, p47). In some preliminary experiments, metabolite peaks co-eluted with one internal standard, which necessitated the use of the other internal standard for subsequent incubations. The human liver microsomes incubation with testosterone and propoxyphene, norpropoxyphene, or proadifen often resulted in peaks that co-eluted with 6β-hydroxy testosterone and/or both internal standards. The mobile phase was adjusted by increasing the percentage of ammonium acetate and decreasing the percentage of methanol to improve separation, although this increased the run time. Subsequently, the pH of ammonium acetate was adjusted to improve separation and reduce run time. A final mobile phase of 50% ammonium acetate pH 5.6-5.8 and 50% methanol was used for human liver microsomes incubations to improve peak separation (Figure 2, p48).

Irreversible Enzyme Inhibition Data:

Propoxyphene, norpropoxyphene, and proadifen exhibited time and concentration dependent inactivation of CYP3A4(+b5), CYP3A5(+b5), and CYP3A with human liver microsomes (see Table 4, p40). The lowest KI was achieved by proadifen and

CYP3A4(+b5) (0.35 µM). Propoxyphene and human liver microsomes generated the second lowest KI (0.45 µM). All three compounds were less potent inhibitors of

CYP3A5(+b5) than CYP3A4(+b5) based on KI values. The KI values with CYP3A5(+b5) were three to sixty times higher than corresponding values with CYP3A4(+b5) and three to thirty times higher than with human liver microsomes. The KI values for propoxyphene with human liver microsomes were lower than the KI values of propoxyphene with CYP3A4(+b5) and CYP3A5(+b5). Norpropoxyphene exhibited inactivation (KI values) with CYP3A4(+b5) and human liver microsomes at similar concentrations. Proadifen was a more potent inhibitor of CYP3A4(+b5) than human liver microsomes (KI values were 0.35 and 6.9 µM, respectively).

Proadifen exhibited irreversible inhibition of CYP3A4(+b5), CYP3A(+b5), and

CYP3A in human liver microsomes. The KI values for proadifen and CYP3A5(+b5) were approximately three-fold higher than the KI values for proadifen and human liver microsomes. The rates of enzyme inactivation are summarized in Table 4, p40.

Propoxyphene and norpropoxyphene exhibited the highest rates of inactivation (highest

-1 -1 kinact) with CYP3A4(+b5), 0.41 min and 0.56 min , respectively. The highest kinact of all three compounds with CYP3A5(+b5) was achieved by norpropoxyphene (0.21 min-1).

The rates of inactivation for proadifen and CYP3A4(+b5) and proadifen and human liver microsomes were similar (0.26 min-1 and 0.20 min-1, respectively).

The results of irreversible inhibition experiments are plotted in Figures 3-29 (p49-

75). These include graphs of % Activity v. Pre-incubation Time, kobs v. Inhibitor, and %

Activity Relative to Control. The percent activity versus pre-incubation time graphs include the averaged data points and lines calculated from model estimates, these are listed in Figures 3-11 (p49-57). Although the pre-incubation times differ, the plots demonstrate that activity decreases as pre-incubation time increases, and activity

decreases with increasing concentrations of inhibitor. The lines in Figures 3-11 were

calculated from Equation 2b (p46) with kinact and KI model estimates. Although not all %

activity data points fall on the line, the general trend of the data for each concentration is

similar to the predicted line shown in the graph. The slopes of excel best fit lines from

Figures 3-11 which represent kobserved, were plotted versus inhibitor concentration in

Figures 12-20 (p58-66). The hyperbolic curve was calculated (Equation 2a, p46) based

on kinact and KI model estimates using percent activity.

Figures 21-29 (p67-75) show percent of control enzyme activity remaining using

pre-incubation for control or inhibitor reactions. The concentrations of inhibitor are listed

across the x-axis as inhibitor concentration in the total incubation mix (1 mL reaction)10.

Only norpropoxyphene and CYP3A5(+b5) showed a decrease in % activity relative to control below 75% in these experiments performed without pre-incubation

(approximately 70% activity at 4 µM norpropoxyphene, Figure 25, p71).

Metabolic-Inhibitor Complex Formation Data:

Propoxyphene, norpropoxyphene, and proadifen formed metabolic-inhibitor complexes with CYP3A4(+b5) and CYP3A5(+b5) (see Figure 30, p76 for a metabolic- inhibitor complex plot, and Table 4, p40 for metabolic-inhibitor complex formation results for each compound and enzyme). In initial studies with the dual beam spectrophotometer and CYP3A5(+b5) enzyme, only propoxyphene and proadifen formed a metabolic-inhibitor complex with CYP3A5(+b5). Norpropoxyphene had been tested for metabolic-inhibitor complex formation with CYP3A5(+b5) before the KI was determined

by pre-incubation experiments, and the concentration may have been too low to form a

10Although this control is not pre-incubated with NADPH prior to the addition of substrate, the total concentration of inhibitor in a 50 µL pre-incubation reaction is listed in parentheses, for ease of comparison to pre-incubated samples.

metabolic-inhibitor complex (norpropoxyphene was tested at 15.3 µM, the KI was later determined to be 25.2 µM). In follow up experiments with the plate reader, norpropoxyphene was tested at a higher concentration (153 µM) and formed a metabolic- inhibitor complex with CYP3A5(+b5). Propoxyphene and norpropoxyphene did not form a metabolic-inhibitor complex with human liver microsomes, although proadifen did form a metabolic-inhibitor complex with human liver microsomes.

Reversible Enzyme Inhibition Data:

Propoxyphene, norpropoxyphene, and proadifen exhibited reversible inhibition of

CYP3A4(+b5), but proadifen was the most potent inhibitor (Ki value was 5 µM) (Table

5, p41). Propoxyphene and norpropoxyphene exhibited reversible inhibition of

CYP3A4(+b5) only at higher concentrations (Ki values were 26 µM and 29 µM, respectively). Proadifen was the most potent reversible inhibitor of CYP3A5(+b5) and human liver microsomes, with Ki values of 12 µM and 8 µM, repectively.

Norpropoxyphene exhibited reversible inhibition of human liver microsomes (Ki value was 59 µM), but the estimated Ki value of propoxyphene (155 µM) was higher than the highest concentration tested (80µM). The Ki estimates for propoxyphene or norpropoxyphene and CYP3A5(+b5) were also above the highest concentration tested

(100 µM). For propoxyphene and CYP3A4(+b5), the competitive model had the lowest

SE, CV, AIC, and SBC. For all compounds and enzymes studied, the competitive model yielded the best fit of the data and the lowest overall values for SE, CV, AIC, and SBC.

A Comparison of Irreversible and Reversible Enzyme Inhibition Data:

Propropoxyphene and norpropoxyphene were more potent irreversible inhibitors than reversible inhibitors. The Ki value for propoxyphene and CYP3A4(+b5) is twenty

times higher than the KI value, and the Ki value for propoxyphene and human liver microsomes is over one hundred and eighty times higher than the KI value. For propoxyphene and CYP3A5(+b5), the Ki value is more than seven times higher than the

KI value For norpropoxyphene, the Ki value of CYP3A4(+b5) was three times higher than KI values, and the Ki value for human liver microsomes was seven times higher than the KI value. Although the Ki value for norpropoxyphene could not be determined, the Ki value is greater than four times the value of KI value. Unlike propoxyphene and norpropoxyphene, proadifen achieved similar concentrations for Ki value and KI value with human liver microsomes and CYP3A5(+b5). The Ki value and KI value for proadifen and human liver microsomes are 8 µM and 7 µM, respectively. The Ki value and KI value for proadifen and CYP3A5(+b5) are 20 µM and 12 µM, respectively.

Irreversible inhibition (KI value) of CYP3A4(+b5) was ten times greater than reversible inhibition (Ki value) with proadifen, KI value was 0.4 µM and Ki value was 5 µM (Tables

4 and 5, p40-41). The Ki values of proadifen with CYP3A5(+b5) and human liver microsomes are twelve and eight, respectively, and KI values are twenty and seven, respectively. The Ki value for proadifen and CYP3A4(+b5) is about fourteen times the KI value.

CHAPTER 4-DISCUSSION

The results are discussed relative to the distribution of propoxyphene in the body, enzyme selectivity, reversible and irreversible inhibition, in vivo propoxyphene and norpropoxyphene concentrations, and drug-drug interactions with other CYP3A substrates. Proadifen results will also be mentioned.

After oral administration, propoxyphene is rapidly distributed to the liver, brain, lungs, and kidneys, and is eliminated as propoxyphene or metabolized product

(norpropoxyphene or dinorpropoxyphene) in the urine (Clark, 1986). However, orally- administered propoxyphene must pass through the small intestine and liver before distributing to the rest of the body11, and is therefore subject to “first-pass” metabolism

(Ferrier, 1972). Only eighteen percent of propoxyphene enters the systemic circulation from the oral administration of a 65 mg dose of propoxyphene hydrochloride (Ferrier,

1972). Propoxyphene and norpropoxyphene can be metabolized by CYP3A in the small intestine and liver to dinorpropoxyphene, and/or be irreversibly bound to CYP3A enzymes. CYP3A4 is the most abundant CYP in the liver and small intestine, where

CYP3A5 is also found (Shimada et al, 1994), although functional protein is only expressed in some individuals (Kamden et al, 2005; Huang et al, 2004). Propoxyphene or norpropoxyphene that reaches the systemic circulation may be metabolized by CYP3A5 in the kidney (Eichelman and Burk, 2001). Selectivity of CYP3A5 over CYP3A4 would be important to know, as CYP3A4 and CYP3A5 vary in interpersonal expression of functional enzyme and distribution in the body.

11 The drug is most often taken orally because intravenous and subcutaneous administration result in severe vein and soft tissue damage (Hudson, 1977). Oral administration is also more convenient than i.v.

Propoxyphene is a more potent mechanism-based inhibitor of CYP3A4 than of

CYP3A5, with KI’s approximately ten times higher for CYP3A4 than CYP3A5 (Table 4, p40). In fact, propoxyphene, norpropoxyphene, and even proadifen were more potent irreversible inhibitors of CYP3A4 than CYP3A5 based on KI’s. CYP3A5 generally exhibits a lower metabolic capability than CYP3A4 (Williams et al, 2002), but there are exceptions, such as vincristine and tacrolimus, which are metabolized by CYP3A4 and

CYP3A5 with equal efficiency12 (Dennison et al, 2007; Kamden et al, 2005).

Compounds exhibiting higher catalytic rates of metabolism with CYP3A5 than with

CYP3A4 may show increased toxicity in patients who do not express functional

CYP3A5. Additionally these compounds may exhibit reduced efficacy due to lower blood levels in patients who express high levels of functional CYP3A5. The clearance rates of tacrolimus and vincristine increased in CYP3A5 high expressers as compared with low expressers (McPhee et al, 2002; Dennison et al, 2007). Increased adverse events have been associated in cancer patients taking vincristine who have low CYP3A5- mediated metabolism as compared to patients with functional CYP3A5 (Dennison et al,

2007).

Patients who express functional CYP3A5 and CYP3A4 may exhibit different drug plasma concentrations and clearance rates than patients who do not express functional CYP3A5, depending on the drug. Individuals expressing functional CYP3A5

(and CYP3A4) possess two enzymes capable of metabolizing one substrate. Chow et al

(2006) found that poor CYP3A5 expressers (CYP3A5*3/*3) had higher propoxyphene plasma concentrations and lower propoxyphene clearance rates than high expressers

12 The metabolic capability or efficiency of isozymes can be compared by clearance rates (Williams et al, 2002).

(CYP3A5*1/*3 or CYP3A5*1/*1). It is reasonable to conclude that CYP3A5 high expressers (CYP3A5*1/*1) may be less susceptible to propoxyphene-related adverse events, despite propoxyphene’s selectivity for CYP3A4 over CYP3A5.

Polymorphisms in CYP3A4 may also affect the clearance and concentration of

CYP3A substrates. In vivo metabolism of 3A4 substrates may vary up to ten-fold because of differences in CYP3A4 expression (Danielson, 2002). CYP3A4*17, CYP3A4*18, and

CYP3A4*1B alleles are associated with differential CYP3A4 expression as compared to wildtype (CYP3A4*1, and propoxyphene and norpropoxyphene concentrations may vary in individuals based on CYP3A4 genotype (Dai et al, 2001; Kuehl, 2001). The

CYP3A4*17 genotype corresponds with reduced CYP3A4 expression. Accordingly,

AUCs for propoxyphene and norpropoxyphene may be higher in individuals that have this genotype, and they may experience an increased incidence of adverse events.

Because CYP3A expression and metabolism studies with CYP3A4*1B alleles have produced contradictory results, it is difficult to predict the extent of drug interactions with this genotype. Findings include CYP3A4*1B producing two-fold increased CYP3A activity over wildtype (Kadlubar, 2003), no change in CYP3A4 expression with

CYP3A4*1B alleles, and no change in midazolam clearance as compared to wildtype

CYP3A4*1 (Rebbeck, 2000). The CYP3A*1B allele is in linkage disequilibrium with the

CYP3A5 high expression allele (CYP3A5*1), and high CYP3A5 expression may confound CYP3A drug metabolism results.13 Propoxyphene and norpropoxyphene blood concentrations may be lower in individuals with the CYP3A4*18 genotype, which corresponds with increased CYP3A4 expression (Dai et al, 2001).

13 The non random association of genes at more than one loci. 80% of Caucasians with CYP3A4*1B allele also possessed one CYP3A5*1A allele (Wojoski et al, 2002).

Despite the existence of CYP3A4 and CYP3A5 polymorphisms in the general

population, one would expect in vivo studies to support the in vitro observations that

propoxyphene inhibits CYP3A metabolism. The studies by Inturrisi et al (1982) support

the assertion that propoxyphene (and norpropoxyphene) are possibly mechanism-based

inhibitors of CYP3A. Inturrisi et al (1982) reported that repeated dosing of propoxyphene

resulted in the accumulation of propoxyphene and norpropoxyphene in patients, and

blood concentrations of propoxyphene and norpropoxyphene were five to seven times

higher than the concentrations achieved after a single dose. They also found that the

clearance of propoxyphene and norpropoxyphene decreased with repeated dosing (994 to

508 mL/min and 454 to 210 mL/min, respectively). Inturrisi et al (1982) found that the

half life of the two compounds increased from 3.3 to 11.8 hours for propoxyphene, and

from 6.1 to 39.2 hours for norpropoxyphene with repeated dosing. These characteristics

may be attributed to a mechanism-based inhibitor. As CYP3A enzymes are irreversibly

inhibited by propoxyphene and norpropoxyphene, the free enzyme pool is depleted, and

these compounds may accumulate.

Propoxyphene, norpropoxyphene, and proadifen are mechanism-based inhibitors of CYP3A as measured by inhibition of CYP3A4(+b5), CYP3A5(+b5), and CYP3A in human liver microsomes. Propoxyphene and norpropoxyphene form metabolic-inhibitor complexes with CYP3A4(+b5) and CYP3A5(+b5), but proadifen formed metabolic- inhibitor complexes with the recombinant CYPs and human liver microsomes. Ernest et al (2004) observed similar results with protease inhibitors; metabolic-inhibitor complex formation occurred with protease inhibitors and CYP3A4(+b5), but not with human liver microsomes. The pooled liver microsomes express CYP3A4 and CYP3A5 (Wang, 2005),

but the lack of metabolic-inhibitor complex formation could be attributed to the low amount of CYP3A compared to total protein in human liver microsomes. The pooled livers contained approximately three hundred pmoles of CYP (Gorski et al, 1994), and approximately thirty pmoles of CYP3A enzyme per mg of total protein (Wang et al,

2005). The concentration of CYP3A in human liver microsomes is near the limit of quantitation of twenty-three pmoles (Ernest, 2004). Proadifen is a general CYP inhibitor

(Bensoussan et al, 1995) and may bind irreversibly to multiple CYPs, therefore it is not surprising that it forms a metabolic-inhibitor complex with human liver microsomes. The binding spectra (450 nm) of other CYPs may resemble the binding spectra of CYP3A.

CYP2D6 and paroxetine form a metabolic-inhibitor complex at approximately 450 nm

(Bertelsen, 2003). Additionally, proadifen forms a metabolic-intermediate-complex with guinea pig CYP2B6 which may resemble the CYP3A4 spectra (Yamada et al, 1992).

Therefore it is reasonable that proadifen would form a metabolic-inhibitor complex with human liver microsomes, and propoxyphene and norpropoxyphene did not form a metabolic-inhibitor complex with human liver microsomes.

Propoxyphene and norpropoxyphene are weak reversible inhibitors of CYP3A, with KI values greater than or equal to 26 µM. Many Ki value estimates were greater than the highest concentrations tested (≥ 80 µM), much higher than drug concentrations in plasma (propoxyphene and CYP3A5(+b5), norpropoxyphene and CYP3A5(+b5), and propoxyphene and human liver microsomes).

Propoxyphene and norpropoxyphene were more potent irreversible inhibitors than reversible inhibitors of CYP3A4(+b5), CYP3A5(+b5), and CYP3A in human liver microsomes. Because irreversible inhibition removes functional enzyme from the enzyme

pool, much of the CYP3A functional protein would theoretically be removed. The KI values (inhibitor concentration at half kinact) were at least three times lower than the Ki values (concentration of enzyme-inhibitor relative to free enzyme and free inhibitor) for

CYP3A4(+b5), CYP3A5(+b5), and human liver microsomes for propoxyphene and norpropoxyphene. The reported therapeutic blood concentrations of propoxyphene are listed in Table 6 (p42) and range from 0.4 to 2.5 µM. The KI value for propoxyphene (~1

µM) is within range of reported therapeutic blood levels, whereas the Ki value is much higher (~50 µM). Therefore propoxyphene concentrations in the body would not approach the levels needed for reversible inhibition based on in vitro data. The toxic blood concentrations of propoxyphene are greater than 1.5 µM, which closely corresponds to the KI value for CYP3A4 (Merck, 2007). The reported therapeutic concentrations of norpropoxyphene in blood range from 0.9-15 µM (Verbeley and

Inturrissi, 1973; Inturrisi et al, 1982). These concentrations are within the range of KI value for norpropoxyphene (~8 µM), whereas the Ki value for norpropoxyphene is approximately 40 µM. Norpropoxyphene blood concentrations would not approach the levels needed for reversible inhibition based on in vitro data. In vivo inhibition of CYP3A enzymes would probably be due to mechanism-based inhibition and not due to reversible inhibition.

Norpropoxyphene may play a significant role in adverse events attributed to propoxyphene because it has potent anesthetic properties and causes cardiac toxicity and seizures (Nickander et al, 1984). Approximately seventy-six percent of propoxyphene deaths are attributed to cardiac toxicity (Whitcomb et al, 1989). Norproxyphene causes hypotension, decreased contractability, and interruption of cardiac conduction (Holland

and Steinberg, 1979). Propoxyphene and norpropoxyphene have anti-arrhythmic

properties and block sodium channels, but norpropoxyphene is more potent than

propoxyphene for cardiac effects (Holland and Steinberg, 1979; Slywka, 1975).

Norpropoxyphene also has a longer half life than propoxyphene (30-36 hours versus 6-12

hours), and can accumulate in the body. Norpropoxyphene blood concentrations as high

as 15 µM have been found after high oral therapeutic doses of propoxyphene (Inturrisi et al, 1982). Toxicity for norpropoxyphene has not been established, but toxicity has been associated with blood concentrations of 0.15 µM propoxyphene (see Table 6, p42).

Inhibition of CYP3A by propoxyphene may cause adverse effects in patients

taking high doses of this drug. CYP3A inhibition of propoxyphene may also result in

adverse affects in patients concomitantly administered other CYP3A substrates. The use

of propoxyphene in elderly patients is limited because of the high number of reported

adverse events with propoxyphene in this subpopulation (Beers, 1997), who are often

administered multiple drugs over the same time period (polypharmacy) as compared to

other subpopulations14. Potential drug-drug interactions exist for propoxyphene and other

CYP3A substrates. Mechanism-based inhibition of CYP3A by propoxyphene may

increase the concentrations of other CYP3A substrates in the body.15 CYP3A metabolizes

up to 60% of marketed drugs including some immunosuppressants, heart medications,

and many other drugs (Turgeon et al, 1992; Wang et al, 2005). Other CYP3A substrates

include food products such as caffeine (Tassaneeyakul et al, 1993), and grapefruit juice

(Bailey et al, 1993). Increased plasma concentrations of CYP3A substrates (drugs or

14 The elderly also often have decreased liver function compared to the general population. 15 Some drugs are metabolized by multiple CYP isoforms, and co-administration with propoxyphene may not result in higher blood levels. An example is acetaminophen, which is metabolized to N-acetyl-p- benzoquinone imine by CYP2E1, CYP1A2, CYP2A6, CYP2D6, and CYP3A4 (Dong et al, 2001).

metabolites) can elicit severe adverse events, such as rhabdomyolysis, with high plasma levels of HMG-CoA reductase inhibitors (Dresser et al, 2000) and sedation with benzodiazepines (AHFS, 2007).

It is difficult to determine if inhibition of CYP3A by propoxyphene and norpropoxyphene has contributed to reported adverse events. Accidental and suicide deaths have been attributed to propoxyphene and its metabolite in combination with other medications, including benzodiazepines and other analgesics (McBay, 1976). Several benzodiazepines (alprazolam, diazepam, midazolam, triazolam) and analgesics (cocaine, codeine, fentanyl) are metabolized by CYP3A (Gasche, 2004). Many cases of accidental overdoses with propoxyphene have also occurred in combination with other pain medication (Gram, 1979) such as opioids (Ng and Alvear, 1993), and many opioids are metabolized by CYP3A (Moody, 1996).

Table 7 (p43) lists predicted increases in drug concentrations for some common

CYP3A substrates as a result of drug interactions with propoxyphene and norpropoxyphene through inhibition of CYP3A enzymes. The AUC'po/AUCpo is a ratio of the area under the plasma concentration versus time curve of an orally administered drug in the presence of inhibitor (AUC'po) to the AUC without inhibitor present (AUCpo). The predicted drug interactions (AUC'po/AUCpo’s) with propoxyphene and norpropoxyphene were determined for common CYP3A substrates using Equation 6 (p46). The total blood concentrations of propoxyphene and norpropoxyphene (I), were used to calculate the free concentrations (Iu) based on 76.5% plasma protein binding (from an average of 73-80% plasma protein binding based on Giacomini et al, 1980). The KI and kinact kinetic parameters for propoxyphene and norpropoxyphene and human liver microsomes were

incorporated into Equation 6. The estimates assume a maximal inhibition of intestinal wall CYP3A by propoxyphene and norpropoxyphene (F’G = 1), after repeated dosing.

The estimates of in vivo drug interactions were calculated for low, moderate, and high blood concentrations of propoxyphene and corresponding concentrations of norpropoxyphene, for intravenous midazolam only. Subsequent estimates use only the median concentrations for propoxyphene and norpropoxyphene. The AUC'po/AUCpo values were calculated separately for propoxyphene and norpropoxyphene and then added together for net effect (see Table 7, p43), as per Wang et al’s studies with verapamil and metabolites (2004).

The predicted AUC'po/AUCpo values of orally administered midazolam with propoxyphene are approximately fifteen to twenty-five times the blood concentrations of midazolam administered alone. In general, the lower the intestinal availability of the substrate prior to the addition of inhibitor (FG) and the greater the fraction metabolized by

CYP3A (fm), the greater the predicted change with propoxyphene co-administration.

Sildenafil, triazolam, and R-verapamil have a predicted blood concentrations (AUC) that are at least ten times higher when administered with propoxyphene. Verapamil is a weak

CYP3A inducer, and actual blood levels may be slightly lower than predicted (Wang,

2005). All drugs show a predicted increase in AUC by at least four-fold with propoxyphene (and norpropoxyphene).

Although AUC'po /AUC po data were not available for many of the drugs listed in

Table 7 (p43), some AUC'po/AUCpo values with propoxyphene have been documented.

Abernethy et al (1985) reported that the AUC'po/AUCpo values for alprazolam was 1.6 following three 60 mg propoxyphene doses/day for two days. Using Equation 6, the

predicted AUC'po/AUCpo for alprazolam in humans with co-administration of propoxyphene is seven to ten, which is more than three times the reported ratio. The discrepancy between predicted and actual AUC'po/AUCpo for alprazolam may be attributed to a lower frequency of propoxyphene administration compared to the dose used for calculations in Table 7 (p43)16. Additionally, the discrepancy between the predicted and actual AUC'po/AUCpo for alprazolam may be attributed to its low rate of metabolism by the intestine (Obach et al, 2006). If the drug is not as affected by propoxyphene-mediated intestinal CYP inhibition, the actual intensity of the drug interaction may be less than predicted by Equation 6 (AUC'po/AUCpo).

Equation 6 may be more accurate for midazolam and other CYP3A substrates that are metabolized by both the liver and small intestine (Obach et al, 2006). The predicted

AUC changes for oral midazolam with propoxyphene co-administration are fifteen to twenty-five times the AUCpo of midazolam alone. Also, midazolam is not transported by

P-glycoprotein, a transport pump which could reduce the amounts of drug in the intestine.

Reducing the amount of drug in the intestine could directly affect the amount of CYP inhibited. (Wang et al, 2004). Abernethy et al (1985) also reported that AUC changes were not observed for diazepam or lorazepam when co-administered with propoxyphene.

These benzodiazepines are not CYP3A substrates as is the case for midazolam, triazolam, and alprazolam.

Propoxyphene has reduced the clearance and increased the half-life of other

CYP3A substrates not included in Table 7 (p43) due to lack of sufficient information for

16 Test subjects taking 65 mg propoxyphene hydrochloride every 6 hours for two and a half-days were administered 1.0 mg of alprazolam once (Abernethy, 1985). Table 6 AUC changes are based on a plasma concentration of 0.6 µM propoxyphene from administration of 65 mg propoxyphene hydrochloride three times a day for four days (Verbeley and Inturrisi, 1973). See Table 5.

calculations with Equation 617. Propoxyphene decreased the total metabolic clearance of antipyrine from 0.53 to 0.63 mL/min, resulting in an increase in the elimination half-life of antipyrine from 12.2 to 15.2 hours (Abernethy, 1982). Abernethy et al (1982) found that propoxyphene increased the steady-state plasma levels of doxepin and desmethyldoxepin from 19 to 44 ng/mL and 9 to 20 ng/mL, depressing cognitive function proportionally.

A drug-drug interaction can have serious clinical consequences if the difference between toxic and effective concentrations is small (Lin and Lu, 1998), which is the case for propoxyphene (Inturrisi et al, 1982), see Table 6 (p42). Propoxyphene has been associated with many accidental deaths. The high reported rates of accidental overdose may be attributed to the irreversible inhibition of CYP3A enzymes by propoxyphene and norpropoxyphene. The potential drug-drug interactions of propoxyphene through CYP3A inhibition may also have contributed to propoxyphene-related overdoses and adverse events. Inhibition of CYP3A enzymes by propoxyphene and norpropoxyphene may result in higher concentrations of other CYP3A substrates, resulting in adverse events.

Proadifen is not available as a drug but could be used as a positive control for reversible and irreversible inhibition assays with CYP3A. It has been employed in mechanism-based inhibition and metabolic-inhibitor complex formation assays with

CYP3A and other CYPs (Yamada et al, 1992; Jones et al, 2007). Proadifen is more potent than propoxyphene and norpropoxyphene for irreversible inhibition with

CYP3A4(+b5). It is also a more potent reversible inhibitor than propoxyphene and norpropoxyphene with CYP3A4(+b5), CYP3A5(+b5), and human liver microsomes. It is

17 The fraction of the total hepatic elimination due to CYP3A in the absense of inhibitor (fm) and the intestinal wall bioavailability of the substrate in the absense of inhibitor (FG) could not be obtained from published literature.

not a controlled substance like propoxyphene (requiring less paperwork and control for laboratory use).

CHAPTER 5-CONCLUSION

Propoxyphene was developed over fifty years ago, before the current knowledge of CYP isozymes, metabolism, and CYP-mediated drug-drug interactions. The results of these studies show that propoxyphene and norpropoxyphene are irreversible inhibitors of

CYP3A as measured by in vitro experiments with CYP3A4(+b5), CYP3A5(+b5), and human liver microsomes. Propoxyphene and norpropoxyphene exhibit little or no reversible inhibition of CYP3A, and are more potent irreversible inhibitors of CYP3A.

Both propoxyphene and norpropoxyphene form metabolic-inhibitor complexes with

CYP3A4(+b5) and CYP3A5(+b5). Proadifen, a compound of similar structure to propoxyphene and norpropoxyphene, is a potent irreversible inhibitor of CYP3A4.

Proadifen is also an irreversible inhibitor of CYP3A5 and human liver microsomes, and exhibits reversible inhibition with CYP3A4(+b5), CYP3A5(+b5), and human liver microsomes.

Many reported propoxyphene overdoses may be accidental and attributed to the irreversible inhibition of CYP3A enzymes by propoxyphene, which may result in higher than predicted blood concentrations of the drug or its metabolite, norpropoxyphene.

Inhibition of CYP3A enzymes by propoxyphene and norpropoxyphene may result in higher concentrations of other CYP3A substrates, resulting in adverse events.

Future studies may be conducted to provide additional information regarding propoxyphene and norpropoxyphene inhibition of CYP3A. These include testing

CYP3A4(+b5) and CYP3A5(+b5) for regeneration of activity following propoxyphene and norpropoxyphene pre-incubation experiments to determine if the irreversible inhibition is completely irreversible. Additionally, propoxyphene and norpropoxyphene

may be examined in combination in irreversible inhibition experiments to assess cumulative, additive, or synergistic effect. A study may also be conducted to monitor concentrations of propoxyphene, norpropoxyphene and dinorpropoxyphene following incubation with CYP3A to assess depletion and product formation. Additionally, propoxyphene and norproxyphene may be tested for induction of CYP3A enzymes in vitro using established cell culture models. Although propoxyphene and norpropoxyphene exhibited irreversible inhibition with recombinant CYP3A enzymes, the in vivo inhibition may be less than predicted due to CYP3A enzyme induction, which could not be assessed with pre-incubation experiments.

A study may also be conducted to compare theoretical and actual AUC values of patients taking propoxyphene alone and in combination with other CYP3A substrates. As a part of this study, patients may be genotyped to determine if any CYP3A polymorphisms exist. Providing genotyping information would enhance data interpretation, and highlight which sub-populations, if any, are more susceptible to propoxyphene mediated CYP3A inhibition and drug-drug interactions.

Table 1. Chemical Structures of Propoxyphene, Methadone, Codeine, Norpropoxyphene and Dinorpropoxyphene.

CH O N 3

H3C CH3 Propoxyphene CH3 O CH2

CH H C 3 Methadone 3 N O CH3 CH3

N CH3 Codeine

CH O 3 O OH

CH Norpropoxyphene O N 3 H3C H CH3 O CH2

Dinorpropoxyphene O N H H C H 3 CH CH 3 (linear) O 2

(Feinberg et al,1976; McMahon, 1961; Somogyi et al, 2004)

Table 2. Chemical Structures of Proadifen (SKF-525a) and SKF-8742

H3C O O CH Prodifen (SKF-525a) N 3

CH3

H3C O O N CH3 SKF-8742 H

(Anders and Mannering, 1966)

Table 3. List of CYP3A4 and CYP3A5 Common Polymorphisms

Enzyme Expression of Functional Protein in Major Alleles Activity Relative to Wild Type Adult Population Organs CYP3A4 CYP3A4*1(A) Wildtypef

CYP3A4*17 Reduced CYP3A4 expressiona

Liver Majority of population expresses CYP3A4 CYP3A4*18 a Small Intestinef Increased CYP3A4 expression 38 CYP3A4*1B Increased CYP3A4 activity over wildtypeb Decreased CYP3A4 expression over wildtypec

CYP3A5*1(A) Wildtype, Kidney Dominant Allele, produced functional protein (*1/*3)e 1-30% of Caucasiansd CYP3A5 Lung 55-75% of Black Africans and African Americansd d Liver 33% of Japanese f CYP3A5 *2, Colon Less than 30% of wild type CYP3A5 proteine *3, *5, *6, *7

aDai et al, 2001; bKuehl et al, 2001; cWojnoski et al, 2002; dKamden et al, 2005; eKreutz et al, 2005; fDanielson, 2002

Table 4. Summary of Kinetic Parameters for Enzyme Inactivation (Irreversible Inhibition) with Propoxyphene, Norpropoxyphene, and Proadifen

Tissue Inhibitor Kinact (min-1) KI (µM) MIC Formation

Propoxyphene 0.41 (+/- 0.03) 1.3 (+/- 0.28) Yes CYP3A4(+b5) Norpropoxyphene 0.56 (+/- 0.07) 8.8 (+/- 2.1) Yes Proadifen 0.26 (+/- 0.03) 0.35 (+/- 0.16) Yes Propoxyphene 0.072 (+/- 0.005) 13 (+/- 3.0) Yes CYP3A5(+b5) Norpropoxyphene 0.21 (+/-0.03) 25 (+/-8.7) Yes Proadifen 0.11 (+/-0.01) 20 (+/-4.6) Yes Propoxyphene 0.038 (+/-0.002) 0.45 (+/-0.13) None detected

39 Human Liver Microsomes Norpropoxyphene 0.074(+/-0.004) 8.2 (+/-1.4) None detected Proadifen 0.2 (+/-0.04) 6.9 (+/-2.4) Yes

Table 5. Reversible Inhibition Ki Values

Inhibitor Tissue Ki µM (Standard Error)

CYP3A4(+b5) 26 (+/- 3.0)

Propoxyphene CYP3A5(+b5) >100, estimated 134 (+/-32)

Human Liver Microsomes >80, estimated 155 (+/-22)

40 CYP3A4(+b5) 29 (+/-5.0)

Norpropoxyphene CYP3A5(+b5) >100, estimated 186 (+/-24)

Human Liver Microsomes 59 (+/-0.049)

CYP3A4(+b5) 5.4 (+/-0.042)

Proadifen CYP3A5(+b5) 12 (+/- 1.0)

Human Liver Microsomes 8.3 (+/- 0.61)

Table 6. Reported Therapeutic (Total) Blood Levels of Propoxyphene and Norpropoxyphene

Propoxyphene Propoxyphene Norpropoxyphene Norpropoxyphene Dose Frequency Duration (µg/mL) (µM) (µg/mL) (µM)

130 mg Propoxyphene 1 time only once 0.3a 0.9 0.3 a 0.9 hydrochloride a 65 mg Propoxyphene 3 doses/day 4 days 0.1-0.2a 0.4-0.7 0.6 a 1.8 hydrochloride a

130 mg Propoxyphene a a a 3 doses/day 4 days 0.7-0.9 2-2.5 1.1-1.2 3.4-3.7 hydrochloride 41 550 mg Propoxyphene 2 dose/Day 12 weeks 0.5b 1.5 5 b 15 hydrochloride b Toxic levels c n/a n/a >0.5c >1.5 n/a n/a

Therapeutic n/a n/a ≥ 0.05d ≥ 0.15 n/a n/a concentration d

aVerbeley and Inturrissi, 1973; bInturrisi et al, 1982; c Merck Manual, 2007; dAHFS, 2007

Table 7. Predictions of Propoxyphene and Norproxyphene Interactions with Other CYP3A Substrates (AUC'po/AUCpo)

Propoxyphene Propoxyphene Norproxyphene Propoxyphene Norproxyphene CYP3A + Norproxyphene fm FG Substrates Predicted Predicted Predicted I (µM) Iu (µM) I (µM) Iu (µM) AUC'po/ AUCpo AUC'po/ AUCpo AUC'po/ AUCpo 0.6 a 0.1 1.8 a 0.4 Midazolam-iv c 0.9 c 1 d 5-8 3-6 8-14

2.3a 0.5 3.6 a 0.8 Midazolam-iv c 0.9 c 1 d 7-9 4-8 11-17

1.5b 0.4 15 b 3.5 Midazolam-iv c 0.9 c 1 d 6-9 6-9 12-18 0.6 a 0.1 1.8 a 0.4 Midazolam-oral c 0.9 c 0.4 c 8-10 7-15 15-25 0.6 a 0.1 1.8 a 0.4 Sildenafil e 0.8f 0.4g 8-11 6-10 14-21 0.6 a 0.1 1.8 a 0.4 Alprazolamo 0.8 l 0.9 l 4-5 3-5 7-10

42 a 0.6 0.1 1.8 a 0.4 Triazolam k 0.8 l 0.4l 8-10 6-9 14-19 0.6a 0.1 1.8 a 0.4 Trazodone h 0.4i 0.8 j 2 2 4 0.6 a 0.1 1.8 a 0.4 R-Verapamilm 0.8m 0.5n 5-6 5-9 10-15 0.6 a 0.1 1.8 a 0.4 S-Verapamilm 0.7m 0.5n 7-11 7-11 14-22

aVerbeley and Inturrisi, 1973,; bAHFS, 2007; c Palkama et al, 1999; d Ernest et al, 2004; e Muirhead et al; f Warrington et al, 2000; g Thummel and Shen, 2001;h Greenblatt et al, 2003;i Jaunch et al, 1976;j Nilson and Dale, 1992; kGreenblatt et al, 2000a;l Rodrigues et al, 2001; m Wang et al, 2004; n Gorski et al, 1998; oGreenblatt et al, 2000b

The predicted AUC'po/AUCpo values were determined for propoxyphene and norpropoxyphene separately using Equation 6, and then added together. The total blood concentrations of propoxyphene and norpropoxyphene (I), were used to calculate the free concentrations (Iu) based on 76.5% plasma protein binding (Giacomini et al, 1978). The fraction of total hepatic elimination of substrate due to CYP3A in the absence of inhibitor is fm, and was obtained from the scientific literature. FG is the intestinal wall bioavailability of the substrate in the absence of inhibitor. The endogenous degradation rate of CYP3A (kdeg) were 0.00128 and 0.00026 min-1 based on rat CYP3A and human CYP3A4 in CaCO-2 cells (Correa, 1991; Malhotra et al, 2001).

Schematic I. Irreversible Inhibition

k 1 k 2 k 3 Enzyme + Inhibitor EI Intermediate Enzyme + Metabolite k -1 k 4

Inactivated Enzyme (Silverman, 1995)

Schematic II. Irreversible Inhibition: Formation of Proposed Metabolic-Inhibitor Complex

OH CH3 H H RN RN RN RN H CH3 CH3 H

R O N N N CYP R O H RNH 2+ N RN Fe OH N N (Bensoussan et al, 1995)O

In this schematic, dialkylamine group of the inhibitor is demethylated twice, oxidized, and finally forms a nitroso group (R-N=O). The two free electrons of the nitrogen in the nitroso group binds to the iron of the heme of the prosthetic group of the CYP.

Schematic III. Reversible Inhibition. Enzyme (E) Binds Substrate (S) or Inhibitor (I).

k1 k2 E+S ES E+P k + -1 I

EI (Stryer, 1996)

The k1, k-1, and k2 rates are listed but are not calculated as part of this thesis.

Equation 1. Irreversible Inhibition Equation for Enzyme Activity (at time (t)), Enzyme

Activity at Time (0), and Kobserved (Kobs) at Time (t)

E = E × e−(Kobs)t t 0

Equation 2a. Irreversible Inhibition Equation for kobserved (kobs), kinact, and KI.

k × I k = inact obs K I + I

Equation 2b. Irreversible Inhibition Equation for kobserved, kinact, and KI and Enzyme Activity k ×I −( inact )t K I +I E t = E0 ×e

Equation 3. Competitive Inhibition Equation

Y= (Vmax x S)/((Km x (1+I/ Ki )+S)

Equation 4. Noncompetitive Inhibition Equation

Y= ((Vmax /(1+I/ Ki ))+S)/(Km +S)

Equation 5. Uncompetitive Inhibition Equation

Y= (Vmax x S) /( Km +(S(1+I/ Ki)))

Equation 6. Calculating AUC'po/AUCpo using kinetic parameters

AUC' F' 1 po = G x AUC F f po G m + (1 − f ) k x I m 1 + inact u k deg x (K I + I u )

450.00 1.462 17.160 6β hydroxy testosterone 400.00 Testosterone 350.00

300.00

250.00 Internal Standard mV

46 200.00

150.00

100.00 4.961 50.00 12.736 9.119 0.00 26.312

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30 Minutes

Figure 1. HPLC Chromatogram of Extracted Sample After Incubation with Recombinant CYP. Mobile Phase: 40% 30 mM ammonium acetate pH 6.3-6.4: 60% methanol . HPLC conditions: 5 µm C-18(2) Luna Phenomenex column with a 1 ml/min flow rate and uv detection at 254 nm.

24.00 48.232 22.00 1.465

20.00

18.00 Testosterone

16.00

14.00

12.00 6β hydroxy testosterone Internal Standard

mV 10.00

47 8.00

6.00 33.272

4.00

2.00 9.563

0.00 7.083 11.888 17.600 35.228 -2.00

0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00 Minutes

Figure 2. HPLC Chromatogram of Extracted Sample After Incubation with Human Liver Microsomes. Mobile Phase: 40% 30 mM ammonium acetate pH 5.6-5.8: 60% methanol . HPLC conditions: 5 µm C-18(2) Luna Phenomenex column with a 1 ml/min flow rate and uv detection at 254 nm.

100

0.25 ∝Μ 60 0.5 ∝Μ

% Activity 1 ∝Μ 10 ∝Μ 30 ∝Μ

48 50 Μοδελεδ 0.25 ∝Μ Μοδελεδ 0.5 ∝Μ Μοδελεδ 1 ∝Μ Μοδελεδ 10 ∝Μ Μοδελεδ 30 ∝Μ 40 0.00.51.01.5 Pretime (min)

Figure 3. Propoxyphene and CYP3A4(+b5)-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A4(+b5) activity by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

80 5 µM 12.5 µM 25 µM % Activity 49 50 µM 75 µM 70 100 µM 5 µM Modeled Line 12.5 µM Modeled Line 25 µM Modeled Line 50 µM Modeled Line 75 µM Modeled Line 100 µM Modeled Line 60 0123456 Pretime (min)

Figure 4. Propoxyphene and CYP3A5(+b5)-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A5(+b5) activity by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

0.25 µM 70 1 µM

% Activity 5 µM

50 10 µM 20 µM 60 50 µM 0.25 µM Modeled Line 1 µM Modeled Line 5 µM Modeled Line 10 µM Modeled Line 20 µM Modeled Line 50 µM Modeled Line 50 051015 Pretime (min)

Figure 5. Propoxyphene and Human Liver Microsomes-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A activity in human liver microsomes by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100 90

70 0.3 µM 1 µM 51 3 µM 60 6 µM 10 µM % Activity of 0,0 of Activity % 30 µM 0.3 µM Modeled Data Line 50 1 µM Modeled Data Line 3 µM Modeled Data Line 6 µM Modeled Data Line 10 µM Modeled Data Line 30 µM Modeled Data Line 40 0.00 0.25 0.50 0.75 1.00 1.25 1.50

Pretime (min)

Figure 6. Norpropoxyphene and CYP3A4(+b5)-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A4(+b5) activity by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

5 µM 70 10 µM 20 µM

52 40 µM

% Activity 60 µM 60 80 µM 5 µM Modeled Line 10 uM Modeled Line 20 µM Modeled Line 40 µM Modeled Line 50 60 µM Modeled Line 80 µM Modeled Line

01234 Pretime (min)

Figure 7. Norpropoxyphene and CYP3A5(+b5)-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A5(+b5) activity by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

90 80

60 2 µM 5 µM 10 µM 53 50 50 µM % Activity 75 µM 100 µM 2 µM Modeled Line 40 5 µM Modeled Line 10 µM Modeled Line 50 µM Modeled Line 75 µM Modeled Line 100 µM Modeled Line

30 0 5 10 15 Pretime (min)

Figure 8. Norpropoxyphene and human liver microsomes-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A activity in human liver microsomes by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

0.2 µM 80 0.5 µM 54 1.1 µM 2.3 µM 4.5 µM Log of % Activity 9.1 µM 70 0.2 µM Modeled Line 0.5 µM Modeled Line 1.1 µM Modeled Line 2.3 µM Modeled Line 4.5 µM Modeled Line 9.1 µM Modeled Line 60 0.0 0.5 1.0 1.5

Pretime (min)

Figure 9. Proadifen and CYP3A4(+b5)-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A4(+b5) activity by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

70 9 µM

60 23 µM 32 µM % Activity

55 45 µM 50 91 µM 9 µM Modeled Line 23 µM Modeled Line 32 µM Modeled Line 40 45 µM Modeled Line 91 µM Modeled Line

0246810

Pretime (min)

Figure 10. Proadifen and CYP3A5(+b5)-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A5(+b5) activity by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

100

80 70

50 0.9 µM 2.3 µM 56 40 4.5 µM % Activity 9.1 µM 18 µM 30 0.9 µM Modeled Line 2.3 µM Modeled Line 4.5 µM Modeled Line 9.1 µM Modeled Line 18 µM Modeled Line 20 0246810

Pretime (min)

Figure 11. Proadifen and Human Liver Microsomes-Percent Activity v. Pre-incubation Time. Time and concentration dependent inhibition of CYP3A activity in human liver microsomes by propoxyphene. For each concentration of inhibitor, the remaining testosterone 6β hydroxylase activity is expressed as a percentage relative to control activity. Points are averaged data, and lines are generated from Equation 2b (p46) with kinact and KI model estimates.

0.4

0.3

0.2

kobs (1/min) 57

0.1 WinNonLin Modeled Kinact KI Kobs from Excel Best Fit

0.0 0 5 10 15 20 25 30 Propoxyphene (µΜ)

Figure 12. kobs v. Inhibitor Concentration for CYP3A4(+b5) and Propoxyphene. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.08

0.06

0.04 58

kobs (1/min) kobs from Excel Best Fit WinNonLin Modeled kinact KI 0.02

0.00 0 20406080100 Propoxyphene (µM)

Figure 13. kobs v. Inhibitor Concentration for CYP3A5(+b5) and Propoxyphene. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.05

0.04

0.03

59 0.02 kobs (1/min) kobs from Excel Best Fit Line

Windows NonLin Modeled kinact, KI

0.01

0.00 0 1020304050

Propoxyphene (µM)

Figure 14. kobs v. Inhibitor Concentration for Human Liver Microsomes and Propoxyphene. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.5

0.4

0.3

0.2 kobs (1/min)

60 Windows Nonlin Modeled kinact, KI kobs from Excel best fit line 0.1

0.0 0 5 10 15 20 25 30

Norpropoxyphene (µM)

Figure 15. kobs v. Inhibitor Concentration for CYP3A4(+b5) and Norpropoxyphene. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.18

0.12

Windows Nonlin Modeled kinact, KI

kobs from Excel best fit line

61 kobs (1/min) kobs

0.06

0.00 0 20406080 Norpropoxyphene (µM)

Figure 16. kobs v. Inhibitor Concentration for CYP3A5(+b5) and Norpropoxyphene. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.08

0.06

0.04

62 kobs (1/min)

0.02 kobs from Excel best fit line Windows Nonlin Modeled kinact, KI

0.00 0 20406080100

Norpropoxyphene (µM)

Figure 17. kobs v. Inhibitor Concentration for Human Liver Microsomes and Norpropoxyphene. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using equation 2a (p46) and the KI and kinact model estimates.

0.30

0.25

0.20

0.15

kobs (1/min) 63 0.10

Windows Nonlin from kinact KI 0.05 kobs from Excel best fit line

0.00 0246810 Proadifen (µM)

Figure 18. kobs v. Inhibitor Concentration for CYP3A4(+b5) and Proadifen. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.10

0.08

0.06

64 0.04

kobs (1/min)

Windows Nonlin from kinact KI Kobs from Excel best fit line 0.02

0.00 0 20406080100 Proadifen (µM)

Figure 19. kobs v. Inhibitor Concentration for CYP3A5(+b5) and Proadifen. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The Kobs line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

0.16

0.12

0.08

65 0.04 (1/min) kobs Windows Nonlin from kinact KI kobs from Excel best fit line

0.00 0 5 10 15 20

Proadifen (uM)

Figure 20. kobs v. Inhibitor Concentration for Human Liver Microsomes and Proadifen. The kobs values were calculated using the slopes of the lines of best fit from percent activity versus pre-incubation time data. The line was calculated using Equation 2a (p46) and the KI and kinact model estimates.

120

100

% Activity

66 40

0 0 0.0125 (0.25) 0.025 (0.5) 0.05 (1) 0.5 (10) 1.5 (30) Propoxyphene (µM)

Figure 21. Propoxyphene and CYP3A4(+b5)-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

% Activity 40

0 0 (0) 0.25 (5) 0.625 (12.5) 1.25 (25) 2.5 (50) 3.75 (75) 5 (100) Propoxyphene (µM)

Figure 22. Propoxyphene and CYP3A5(+b5)-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

% Activity

68 40

0 0 0.0125(0.25) 0.05(1) 0.25(5) 0.5(10) 1(20) 2.5(50)

Propoxyphene (µM)

Figure 23. Propoxyphene and Human Liver Microsomes-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

60 % Activity

69 40

0 0 0.015(0.3) 0.05(1) 0.15(3) 0.3(6) 0.5(10) 1.5(30)

Norpropoxyphene (µM)

Figure 24. Norpropoxyphene and CYP3A4(+b5)-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

60 % Activity

0 0 0.25 (5) 0.5 (10) 1 (20) 2 (40) 3 (60) 4 (80) Norpropoxyphene (µM)

Figure 25. Norpropoxyphene and CYP3A5(+b5)-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

60 % Activity

71 40

0 0 2 (0.1) 5 (0.25) 10 (0.5) 50 (10) 75 (3.75) 100 (5)

Norpropoxyphene (µM)

Figure 26. Norpropoxyphene and Human Liver Microsomes-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

60 72 % Activity 40

0 0 0.012(0.23) 0.02(0.45) 0.6(1.1) 0.12(2.3) 0.23(4.5) 0.5(9)

Figure 27. Proadifen and CYP3A4(+b5)-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

73 % Activity

0 0 0.45 (9) 1.13 (23) 1.6 (32) 2.3 (45) 4.5 (91) 0.5(9)

Proadifen (SKF-525a) (µM)

Figure 28. Proadifen and CYP3A5(+b5)-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

120

100

74 60 Activity %

0 0 0.045(0.9) 0.11(2.3) 0.23(4.6) 0.45(9.1) 0.9(18) Proadifen (SKF-525a) Incubation Conc uM (Preinc Conc)

Figure 29. Proadifen and Human Liver Microsomes-Percent Activity Relative to Control. The percentage of initial testosterone 6β hydroxylase activity for each inhibitor concentration without pre-incubation time (time 0) activity was determined relative to the no inhibitor zero pre-incubation time (0,0 control). Concentrations of inhibitor are listed across the x-axis as inhibitor concentration (1 mL reaction) and pre-incubation concentrations are listed in parentheses (50 µL reaction). Error bars are standard deviations of replicates.

0.02

0.015

0.01

0.005

-0.005 Abs(-BKG)-Abs(490) -0.01

75 -0.015 30 min 15 min 400 407 414 421 428 435

442 2 min 449 456 463 470 477

Wavelength(nm) 484 491 498

2 min 5 min 10 min 15 min 20 min 25 min 30 min 35 min 40 min

Figure 30. Metabolic-Inhibitor Complex Formation by Propoxyphene with CYP3A4(+b5). Metabolic-inhibitor complex formation was detected by a uv spectrophotometer by calculating the absorbance difference between the experimental cuvette and reference cuvette. Both contained CYP3A4(+b5), phosphate buffer, NADPH, and methanol, but only to a reference cuvette contained propoxyphene. Absorbance was measured scanning wavelengths of 400-500 nm over 30 minutes.

APPENDIX-SUMMARY OF STATISTICAL METHODS

AIC The Akaike Information Criteria determines the goodness of fit for models with different parameters (Akaike, 1987). For comparing multiple models with the same variables (such as competitive, noncompetitive, and uncompetitive modes), the smallest AIC is the best fit.

AIC = Nobservations x log (WRSS) + 2 (NParameters)

CV The coefficient of variation is calculated by taking the standard deviation divided by the mean and expressed as a percentage. It is a way of comparing the degree of variation of one series to another, even if the means are different from one another.

CV = (SD/Mean) x 100

Mean Arithmatic average

SBC The Schwartz Bayesian Criterion measures goodness of fit based on maximum likelihood. For comparing models with similar (or no) weighting, the model with the smallest SBC value is the best fit.

SBC = Nobservations x log (WRSS) + Log(Nobservations) x NParameters

SD The standard deviation of the mean, it is a measure of dispersion of a data set from its mean. It is calculated as the square root of the sum (Σ) of all differences from the mean (xi − µ), etc. squared, divided by the number of observations minus one (n-1). It is equal to the square root of the variance. (Σ square root 2 (xi − µ)) /n-1)

SE The standard error of the mean, calculated by the dividing the standard deviation by the square root of the sample size (n).

SE = SD/(√N)

WRSS Weighted residual sum of squares, estimates the variance of the residuals. (Residuals are observed minus predicted values, and predicted values come from model fitting). It is the sum of squared deviations from the mean where the means are the predicted parameter values.

(Pharsight Windows Nonlin 5.0.1)

REFERENCES

Abernethy D, Greenblatt D, Steel K, Shader R (1982) Annals of Internal Medicine 97(2):223-224

Abernethy D, Greenblatt D, Morse D, Shader R (1985) Br J Clin Pharm 19:51-57

Afshari R, Bateman D (2003) Journal of Tox:Clin Tox 41:560

AHFS Drug Information (2007) Bethesda Maryland. American Society of Health-System Pharmacists, Inc. 2155-2156

Akaike H (1987) Psychometricka 52:317-332

Anders M, Mannering G (1966) Mol Pharmacol 2:328

Anttila S, Hukkanen J, Hakkola J, Stjernvall T, Beaune P, Edwards RJ, Boobis AR, Pelkonen O, Raunio H (1997) Am J Respir Cell Mol Biol 16:242-249

Aoyama T, Yarnano S, Waxman DJ, Lapenson DP, Meyer UA, Fischer V, Tyndale R, Inaba T, Kalow W, Gelboin H (1989) JBiol Chem 264:10388-10395

Bailey D, Arnold J, Munoz C, Spence J (1993) Clin Pharmacol Ther 53:637-642

Beaver WT (1988) Am J Med 84 Supplement 5A:3-15

Beers MH (1997) Arch Intern Med 157:1531–1536

Bertelsen K, Venkatakrishnan K, Von Moltke L, Obach R, Greenblatt D (2003) Drug Metab Dispos 31(3):289-93

Bensoussan C, Delaforge M, Mansuy D (1995) Biochem Pharmacol 49(5):591-602

Boberg M, Angerbauer R, Fey P, Kanhai W, Karl W, Kern A, Ploschke J, Radtke M (1997) Drug Metab Dispos 25(3):321-31

Buening M, Franklin M (1974) Drug Metab and Disp 2(4):386-390

Chow M, Yin O, Tomlinson B (2006) Am Soc Clin Pharmacol Ther PII-54-P50

Clark’s Isolation and Identification of Drugs. Pharmaceutical Press, London 1986

Cook L, Toner JJ, Fellows EJ (1954) J Pharmacol Exp Ther 111(2):131-141

Correia M (1991) Methods Enzymol 206:315-325

Crespi C, Penman B (1997) Adv Pharmacol 43:171-88

Dai D, Tang J, Rose R, Hodgson E, Bienstock R, Mohrenweiser H, Goldstein J (2001) J Pharmacol Exp Ther 299:825-31

Danielson PB (2002) Curr Drug Metab 3:561-597

(DAWN) Trends in Drug-Related Emergency Department Visits 1994-2002 At a Glance (2003) Drug Abuse Warning Network

Dennison J, Kulanthaivel P, Barbuch R, Renbarger J, Ehlhardt W, Hall S (2006) Drug Metab Dispos 34:1317-1327

Dennison, J, Jones D, Renbarger J, Hall S (2007) Pharmacol Exp Ther 321(2):553-563

Dong H, Haining, R, Thummel K, Rettie, A, Nelson, S (2000) Drug Metab Disp 28(12):1397-400

Dossing M, Pilsgaard H, Rasmussen B, Poulsen H (1983) Eur J Clin Pharmacol 25:215- 222

Dresser G, Spence J, Bailey D (2000) Clin Pharmacokinet 38(1):41-57

Drug Topics, Drugtopics.com., Top 200 Selling Drugs by Retail Dollars in 2006:http://www.drugtopics.com/drugtopics/data/articlestandard//drugtopics/072007/405 102/article.pdf

Ducharme J, Abdullah S, Wainer I (1996) Chromatogr B Biomed Appl 678(1):113-28

Eichelbaum M, Burk O (2001) Nature Medicine 7(3):285-287

Engel G, Hofmann U, Heidemann H, Cosme J, Eichelbaum M (1996) Clin Pharmacol Ther 59:613-623

Ernest C, Hall S, Jones D (2004) J Pharmacol Exp Ther 312(2):583-591

Feierman D, Lasker J (1996) Drug Metab Dispos 24(9):932-939

Feinberg A, Creese I, Snyder S (1976) Proc Natl Acad Sci USA 73(11):4215-4219

Franklin M (1977) Pharmacol Ther 2:227-245

Gellner K, Eiselt R, Hustert E, Arnold H, Koch I, Haberl M, Deglmann CJ, Burk O, Buntefuss D, Esher S, Bishop C, Koebe HG, Brinkman U, Klenk HP, Kleine K, Meyer UA, Wojnowski L (2001) Pharmacogenetics 11(2):11-121

Giacomini KM, Gibson TP, Levy G (1980) Clin Pharm Ther Exp 27(4):508-514

Goodman and Gilman's The Pharmacological Basis of Therapeutics 9th edition, 1996 (Hardman J, Limbird L editors) McGraw-Hill, New York

Gorski J, Jones D, Wrighton S, Hall S (1994) Biochem Pharmacol 48:173-182

Gorski J, Hall S, Jones D, VandenBranden M, Wrighton SA (1994) Biochem Pharmacol 47(9):1643-1653

Gram L, Schou J, Way W, Heltberg J, Bodin N (1979) Clin Pharmacol Ther, 26(4):473- 482

Greenblatt D, von Moltke L, Harmatz J, Durol A, Daily J, Graf J, Mertzanis P, Hoffman J, Shader R (2000a) J Acquir Immune Defic Syndr 24:129-136

Greenblatt D, von Moltke L, Harmatz J, Durol A, Daily J, Graf J, Mertzanis P, Hoffman J, Shader R (2000b) Clin Pharmacol Ther 67:335-341

Greenblatt D, von Moltke L, Harmatz J, Fogelman S, Chen G, Graf J, Mertzanis P, Byron S, Culm K, Granada B (2003) J Clin Pharmacol 43:414-422

Gruber C, Miller C, Finneran J, Chernish S (1956) J Pharmacol Exp Ther 118:280-285

Gruber C (1977) JAMA 237(25):2734-2735

Guengerich, F (1999) Annu Rev Pharmacol Toxicol 39:1-17

Holland D, Steinberg M (1979) Toxicol Appl Pharmacol 47(1):123-133

Huang W, Lin Y, McConn D, Calamia J, Totah R, Isoherranen N, Glodowski M, Thummel K (2004) Drug Metab Dispos 32:1434-1445

Inturrisi EC, Colburn WA, Verebrey K, Dayton HE, Woody GE, O’Brien CP (1982) Clin Pharmacol Ther 31:157-167

Jaunch R, Kopitar Z, Prox A, Zimmer A (1976) Arzeneim-Forsch 26:2084-2089

Jones D, Ekins S, Li L, Hall S (2007) Drug Metab Dispos 35(9):1466-1475

Jonsson A, Holmgren P, Ahlner J (2004) Forensic Sci Int 143(1):53-59

Kadlubar F, Berkowitz G, Delongchamp R, Wang C, Green B, Tang G, Lamba J, Schuetz E, Wolff M (2003) Cancer Epidemiol Biomarkers Prev 12(4):327-331

Kamden L, Streit F, Zanger U, Brockmoller J, Oellerich M, Armstrong V, Wojnowski L (2005) Clin Chem 51:1374-1381

Kantola T, Kivisto KT, Neuvonen PJ (1998) Clin Pharmacol Ther 64(1):58-65

Katoh M, Nakajima M, Yamazaki H, Yokoi T (2000) Pharm Res 17(10):1189-1197

Keshava C, McCanlies E, Weston A (2004) Am Journal of Epidemiology 160(9):825-841

Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J (2001) Nat Genet 27:383-91

Kioski A, Vuori E, Ojanpera I (2005) Human and Exper Tox 24:389-396

Koley A, Buters J, Robinson R, Markowitz A, Friedman F(1995) J Biol Chem 270(10):5014-5018

Koudriakova T, Iatsimirskaia E, Utkin I, Gangl E, Vouros P, Storozhuk E, Orza D, Marinina J, Gerber N (1998) Drug Metab Dispos 26(6):552-561

Kreutz R, Zuurman M, Kain D, Silke A, Bolbrinker J, de Jong P, Navis G (2005) Pharmacogenetics and Genomics 15(12):831-837

Levy R, Thummel K, Trager W, Hansten P, Eichelbaum M (2000) Metab Drug Interactions. Philadelphia, Pennsylvania. 314

Lin Y, Dowling A, Quigley S, Farin F, Zhang J, Lamba J, Schuetz EG, Thummel K (2002) Mol Pharmacol 62:162-172

Lin J, Lu A (1998) Clin Pharmacokinet 35(5):361-390

Lister S. Suicide painkiller. The Times. January 31, 2005

Lowery O, Rosebrough H, Farr A, Randall R (1951) J Biol Chem 193:265-275

Malhotra S, Schmiedlin-Ren P, Paine M, Criss A, Watkins P (2001) Drug Metab Rev 33:97

McBay A (1976) Clin Chem 22(8):1319-1321

McBay A, Hudson P (1975) Journal American Medical Association 233:1257

McMahon R (1961) J Med Pharma Chem 4(1):67

McMahon R, Sullivan H, Due S (1973) Life Sci 12:463-473

Merck Manual Professional by Lexi-Comp, accessed online 11/25/2007 http://www.merk.com/mmpe/print/lexicomp/propoxyphene.html

Miller J, Robbins E, Meyers D (1963) J Pharma Sci 52:446-451

Miller RR, Feingold A, Paxinos J (1970) Journal of the American Medical Association 213:996-1006

Muirhead G, Wulff M, Fielding A, Kleinermans D, Buss N (2000) Br J Clin Pharmacol 50:99-107

Murray M(1997) Clinical and Experimental Pharmacology and Physiology 24:465-470

Nagata K, Yamazol Y (2002) Drug Metab and Pharmacok 17(3):167-189

Naritomi Y, Teramura Y, Terashita S, Kagayama A (2004) Drug Metab Pharmacokinet 19:55-61

Nash J, Bennett I, Bopp R, Brunson M, Sullivan H (1975) J Pharm Sci 64(3):429-33

Nelson D, Koymans L, Kamataki T, Stegeman J, Feyereisen, R, Waxman D, Waterman M, Gotoh O, Coon M, Estabrook, R, Gunsalus I, Nebert D (1996) Pharmacogenetics 6:1- 42

Neuvonen P, Jalava K (1996) Clin Pharmacol Ther 60(1):54-61

Ng B, Alvear M (1993) Am J Drug Alcohol Abuse 19(2):153-8

Nickander RC, Emmerson JL, Hynes MD, Steinberg MI, Sullivan HR (1984) Hum Toxicol 3 Suppl:13S-36S

Nilson O, Dale O (1992) Pharmacol Toxicol 71:150-153

Niwa T, Yabusaki Y, Honma K, Matsno N, Tatusuta K, Ishibashi F, Katagiri M (1998) Xenobiotica 28(6):539-547

Obach SR, Walsky RL, Venkatakrishnan K, Gaman EA, Houston JB, Tremaine LM (2006) J Pharmacol Ther Exp 316(1):336-348

Palkama V, Ahonen J, Neuvonen P, Olkkola K (1999) Clin Pharmacol Ther 66(1):33-39

Pershing L, Franklin M (1982) Xenobiotica 12:687-699

Physician’s Drug Reference (2000) 54th Edition. Montvale, New Jersey. Medical Economics Company, Inc. 1578-1581

Polasek T, Miners J (2005) Eu J Clin Pharm 62(3):203-208

Prueksaritanont T, Gorham L, Ma B, Liu L, Yu X, Zhao J, Slaughter D, Arison B, Vyas K (1997) Drug Metab Dispos 25(10):1191-1199

Rebbeck TR (2000) J Natl Cancer Inst 92:76

Rodrigues A, Winchell G, Dobrinska M (2001) J Clin Pharmacol 41:368-373

Sheen C, Dillon J, Bateman D, Simpson K, Macdonald T (2002) Q J Med 95(9):609-19

Schenkman J, Wilson B, Cinti D (1972) Biochem Pharmacol 21:2372-2383

Shimada T, Yamazaki H, Mimura M, Inui Y, Guengerich F (1994) J Pharmacol Exp Ther 270:414-423

Silverman RB (1995) Mechanism-based enzyme inactivators. Methods Enzymol 249:240- 283

Singh R, Chang S, Taylor L (1996) Rapid Commun Mass Spectrom 10:1019-1026

Slywka G, Melikian A, Whyatt P, Meyer M (1975) J Clin Pharmacol 15:598-604

Somogyi A, Menelaou A, Fullston S (2004) Xenobiotica 34(10):875-887

Stryer, L (1996) Biochemistry Fourth Edition. Freeman Press, New York.191-198

Sutton D, Butler AM, Nadin L, Murray M (1997) Pharmacol Exp Ther 282(1):294-300

Tassaneeyakul W, Birkett D, Veronese M, McManus M, Tukey R (1993) J Pharmacol Exp Ther 265(1):401-407

Thummel KE and Shen DD (2001) in Goodman and Gilman's The Pharmacological Basis of Therapeutics 10th edition (Hardman JG and Limbird LE eds) 1917-2023, McGraw-Hill, New York.

Thummel K, Wilkinson G (1998) Annu Rev Pharmacol Toxicol 38:389-430

Turgeon D, Normolle D, Leichtman A, Annesley T, Smith D, Watkins P (1992) Clin Pharmacol Ther 52(5):471-8

Venkatakrishnan K, Greenblatt D, von Moltke L, Shader R (1998) J Clin Psychopharmacol 18(3):256

Verbeley K, Inturrisi C (1973) J Chromatogr 75:195-205

von Moltke L, Greenblatt D, Harmatz J, Duan S, Harrel L, Cotreau-Bibbo M, Pritchard G, Wright C, Shader R (1996) J Pharmacol Exp Ther 276(2):370-379

Wang R, Newton D, Scheri T, Lu A (1997) Drug Metab Dispos 25(4):502-507

Wang Y, Jones D, Hall S (2005) Drug Metab Dispos 33:664-671

Wang Y, Jones D, Hall S (2004) Drug Metab Dispos 32(2):1-8

Warrington J, Shader R, von Moltke L, Greenblatt D (2000) Drug Metab Dispos 28(4):392-397

Whitcomb D, Gilliam F, Starmer C, Grant A (1989) J Clin Invest 84:1629-1636

Williams J, Ring B, Cantrell V, Jones D, Eckstein J, Ruterbories K, Hamman M, Hall S, Wrighton S (2002) Drug Metab Dispos 30(8):883-891

Williams P, Cosme J, Vinkovic D, Ward A, Angone H, Day P, Vonrhein C, Tickle I, and Jhoti H (2004) Science 305:683-686

Wojnowski L, Hustert E, Elein K, Goldammer M, Haberl M, Kirchheiner J, Koch I, Klattig J, Zanger U, Brockmoller J (2002) Journal of the National Cancer Inst 94(8):630- 632

Woosley R (1996) Annu Rev Pharmacol Toxicol 36:233-252

Wrighton S, Brian W, Sari M, Iwasaki M, Guengerich F, Raucy J, Molowa D, Branden M (1990) Mol Pharmacol 38:207-213

Wrighton S, Schuetz E, Thummel K, Shen D, Korzekwa K, Watkins P (2000) Drug Metab Rev 32:339–361

Yamada H, Kaneko H, Takeuchi K, Oguri K, Yoshimura H (1992) Arch Biochem Biophys 299(2):248-54

Yamazaki H, Johnson W, Yune-Fang U, Shimada T, Guengerich F (1996) Journal of Biol Chem 271(44):27438-27444

Yasumori T, Nagata K, Yang S, Chen L, Murayama N, Yamazoe Y, Kato R (1993) Pharmacogenetics 3:291-301

Zubay, GL (1998) Biochemistry. William C. Brown, Dubuque, IA.163-173

CURRICULUM VITAE

Anna Ruth Riley

Education:

Master of Science in Pharmacology/Toxicology, Indiana University, Indpls, Indiana

Bachelor of Arts in Biochemistry, Earlham College, Richmond, Indiana.

Professional Experience:

Analytical Chemist, Manufacturing Science and Technology; Materials Science

Physical/Particle Characterization

Eli Lilly and Company, Indianapolis, Indiana. July 2007-present

Analytical Chemist, Quality Control Laboratories

Eli Lilly and Company, Indianapolis, Indiana. March 2004-July 2007

Associate Analytical Chemist, Quality Control Laboratories

Eli Lilly and Company, Indianapolis, Indiana. November 2003-March 2004

Associate Biologist, Infectious Disease Research

Eli Lilly and Company, Indianapolis, Indiana. February 2000-November 2003

Contractor, Kelly Scientific Resources, Contracted to Infectious Disease Research

Eli Lilly and Company, Indianapolis, Indiana. November 1998-Februrary 2000

Contractor Employee, Kelly Scientific Resources, Contracted to Bioanalytical Systems

(BAS), West Lafayette, Indiana. January-August 1998

Laboratory Technician, HIV and Histoplasmosis Laboratories, Infectious Disease

Research, Indiana University Medical Center. June-December 1996, May-December

1997

Summer Intern, Indianapolis Water Company, Indianapolis, Indiana. Summer 1996

Publication:

Gilmour R, Foster J, Sheng Q, McClain J, Riley A, Sun P, Ng W, Yan D, Nicas T, Henry K, Winkler M (2005) J Bacteriol 187(23):8196